Point cloud data sending / receiving method and computer readable storage medium
By using octree and Morton code point cloud compression technology, the encoding and decoding process of point cloud data is optimized, solving the latency and complexity problems of processing large amounts of point cloud data, and realizing efficient point cloud service support for virtual reality and autonomous driving services.
Patent Information
- Application Number
- CN202411394314.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2019-10-07
- Filing Date
- 2020-08-04
- Publication Date
- 2024-11-15
AI Technical Summary
Existing technologies struggle to effectively process large amounts of point cloud data, resulting in high latency and encoding/decoding complexity, which fails to meet the needs of services such as virtual reality, augmented reality, and autonomous driving.
We employ point cloud compression techniques based on octrees and Morton codes, and optimize the encoding and decoding process of point cloud data through geometric and attribute encoding combined with spatial scalability methods.
It achieves efficient processing of point cloud data, reduces latency and encoding/decoding complexity, and provides high-quality point cloud services to support virtual reality and autonomous driving services.
Smart Images

Figure CN118972624A_ABST
Abstract
Description
[0001] This application is a divisional application of the invention patent application with the original application number 202080069413.9 (International application number: PCT / KR2020 / 010283, application date: August 4, 2020, invention name: Point cloud data sending device, point cloud data sending method, point cloud data receiving device and point cloud data receiving method). Technical Field
[0002] Embodiments relate to a method and apparatus for processing point cloud content. Background Art
[0003] Point cloud content is content represented by a point cloud, which is a collection of points belonging to a coordinate system representing a three-dimensional space. Point cloud content can express media configured in three dimensions and is used to provide various services such as virtual reality (VR), augmented reality (AR), mixed reality (MR), and self-driving services. However, tens of thousands to hundreds of thousands of point data are required to represent point cloud content. Therefore, a method for efficiently processing a large amount of point data is needed. Summary of the invention
[0004] Technical issues
[0005] Embodiments provide a device and method for efficiently processing point cloud data. Embodiments provide a method and device for processing point cloud data to solve delay and encoding / decoding complexity.
[0006] The technical scope of the embodiments is not limited to the above-mentioned technical objectives, and can be extended to other technical objectives that can be inferred by those skilled in the art based on the entire content disclosed in this document.
[0007] Technical Solution
[0008] To achieve these objectives and other advantages and in accordance with the purposes of the present disclosure, as embodied and broadly described herein, a method for receiving point cloud data may include: receiving a bitstream containing point cloud data; and decoding the point cloud data.
[0009] In another aspect of the present disclosure, a method for transmitting point cloud data may include: encoding the point cloud data; and transmitting a bitstream including the point cloud data.
[0010] Beneficial Effects
[0011] The apparatus and method according to the embodiments may efficiently process point cloud data.
[0012] The apparatus and method according to the embodiments may provide high-quality point cloud services.
[0013] The apparatus and method according to the embodiments may provide point cloud content for providing general services such as VR services and self-driving services. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated into and constitute a part of this application. The accompanying drawings illustrate embodiments of the present disclosure and together with the description are used to illustrate the principles of the present disclosure. In order to better understand the various embodiments described below, reference should be made to the description of the following embodiments in conjunction with the accompanying drawings. The same reference numerals will be used throughout the drawings to represent the same or similar parts. In the drawings:
[0015] Figure 1 An exemplary point cloud content providing system according to an embodiment is shown.
[0016] Figure 2 is a block diagram illustrating point cloud content providing operations according to an embodiment.
[0017] Figure 3 An exemplary process for capturing point cloud video according to an implementation is shown.
[0018] Figure 4 An exemplary point cloud encoder according to an implementation is shown.
[0019] Figure 5 An example of a voxel according to an embodiment is shown.
[0020] Figure 6 An example of an octree and occupancy code according to an embodiment is shown.
[0021] Figure 7 An example of a neighbor node pattern according to an embodiment is shown.
[0022] Figure 8 An example of point configuration in each LOD according to an embodiment is shown.
[0023] Figure 9 An example of point configuration in each LOD according to an embodiment is shown.
[0024] Figure 10 A point cloud decoder according to an embodiment is shown.
[0025] Figure 11 A point cloud decoder according to an embodiment is shown.
[0026] Figure 12 A transmitting device according to an embodiment is shown.
[0027] Figure 13 A receiving device according to an embodiment is shown.
[0028] Figure 14 An exemplary structure that can be operated in conjunction with a point cloud data transmitting / receiving method / apparatus according to an embodiment is shown.
[0029] Figure 15 The structure of geometry data and attribute data for spatial scalability according to an embodiment is shown.
[0030] Figure 16 The attribute encoding / decoding process according to an embodiment is shown.
[0031] Figure 17 An octree structure of point cloud data according to an embodiment is shown.
[0032] Figure 18 A method of reorganizing points belonging to a Morton code based node according to an embodiment is shown.
[0033] Figure 19 Shown is merging isolated nodes according to an embodiment.
[0034] Figure 20 A PCC data encoder according to an embodiment is shown.
[0035] Figure 21 A PCC data decoder according to an embodiment is shown.
[0036] Figure 22 A bit stream structure of point cloud data according to an embodiment is shown.
[0037] Figure 23 A sequence parameter set according to an embodiment is shown.
[0038] Figure 24 A set of property parameters according to an embodiment is shown.
[0039] Figure 25 A patch parameter set according to an embodiment is shown.
[0040] Figure 26 An attribute slice header according to an embodiment is shown.
[0041] Figure 27 Attribute encoding / decoding according to an embodiment is shown.
[0042] Figure 28 Bounding boxes and nodes are shown according to an embodiment.
[0043] Figure 29 The matching between various nodes and locations according to an embodiment is shown.
[0044] Figure 30 Shown is the selection of a reference node in an octree structure according to an embodiment.
[0045] Figure 31 A PCC data encoder according to an embodiment is shown.
[0046] Figure 32 A PCC data decoder according to an embodiment is shown.
[0047] Figure 33 A sequence parameter set (SPS) according to an embodiment is shown.
[0048] Figure 34 An attribute parameter set (APS) according to an embodiment is shown.
[0049] Figure 35 A tile parameter set (TPS) according to an embodiment is shown.
[0050] Figure 36 An attribute slice header (ASH) according to an embodiment is shown.
[0051] Figure 37 A method of transmitting point cloud data according to an embodiment is shown.
[0052] Figure 38 A method of receiving point cloud data according to an embodiment is shown. DETAILED DESCRIPTION
[0053] Reference will now be made in detail to preferred embodiments of the present disclosure, examples of which are shown in the accompanying drawings. The detailed description given below with reference to the accompanying drawings is intended to illustrate exemplary embodiments of the present disclosure, rather than to illustrate the only embodiments that can be implemented according to the present disclosure. The following detailed description includes specific details in order to provide a thorough understanding of the present disclosure. However, it will be apparent to those skilled in the art that the present disclosure can be practiced without these specific details.
[0054] Although most of the terms used in the present disclosure are selected from common terms widely used in the art, some terms are arbitrarily selected by the applicant and their meanings are explained in detail in the following description as needed. Therefore, the present disclosure should be understood based on the intended meanings of the terms rather than their simple names or meanings.
[0055] Figure 1 An exemplary point cloud content providing system according to an embodiment is shown.
[0056] Figure 1 The point cloud content providing system shown may include a transmitting device 10000 and a receiving device 10004. The transmitting device 10000 and the receiving device 10004 can transmit and receive point cloud data via wired or wireless communication.
[0057] According to an embodiment, the point cloud data sending device 10000 can obtain and process point cloud video (or point cloud content) and send it. According to an embodiment, the sending device 10000 may include a fixed station, a base transceiver system (BTS), a network, an artificial intelligence (AI) device and / or system, a robot, an AR / VR / XR device and / or a server. According to an embodiment, the sending device 10000 may include a device configured to communicate with a base station and / or other wireless devices using a radio access technology (e.g., 5G new RAT (NR), long term evolution (LTE)), a robot, a vehicle, an AR / VR / XR device, a portable device, a home appliance, an Internet of Things (IoT) device, and an AI device / server.
[0058] According to the implementation manner, the sending device 10000 includes a point cloud video acquirer 10001, a point cloud video encoder 10002 and / or a transmitter (or communication module) 10003.
[0059] The point cloud video acquirer 10001 according to the embodiment acquires the point cloud video through a processing process such as capture, synthesis or generation. The point cloud video is a point cloud content represented by a point cloud, which is a set of points located in a 3D space and can be referred to as point cloud video data. The point cloud video according to the embodiment may include one or more frames. A frame represents a still image / picture. Therefore, the point cloud video may include a point cloud image / frame / picture and may be referred to as a point cloud image, frame or picture.
[0060] The point cloud video encoder 10002 according to the embodiment encodes the acquired point cloud video data. The point cloud video encoder 10002 may encode the point cloud video data based on point cloud compression coding. The point cloud compression coding according to the embodiment may include geometry-based point cloud compression (G-PCC) coding and / or video-based point cloud compression (V-PCC) coding or next generation coding. The point cloud compression coding according to the embodiment is not limited to the above-mentioned embodiment. The point cloud video encoder 10002 may output a bit stream containing the encoded point cloud video data. The bit stream may include not only the encoded point cloud video data, but also signaling information related to the encoding of the point cloud video data.
[0061] The transmitter 10003 according to the embodiment transmits a bitstream containing encoded point cloud video data. The bitstream according to the embodiment is encapsulated in a file or a fragment (e.g., a stream fragment) and is transmitted via various networks such as a broadcast network and / or a broadband network. Although not shown in the figure, the transmitting device 10000 may include a packager (or a package module) configured to perform a packaging operation. According to the embodiment, the packager may be included in the transmitter 10003. According to the embodiment, the file or fragment may be sent to the receiving device 10004 via a network, or stored in a digital storage medium (e.g., USB, SD, CD, DVD, Blu-ray, HDD, SSD, etc.). The transmitter 10003 according to the embodiment can communicate with the receiving device 10004 (or the receiver 10005) via a 4G, 5G, 6G or other network by wire / wireless communication. In addition, the transmitter may perform necessary data processing operations according to the network system (e.g., a 4G, 5G or 6G communication network system). The transmitting device 10000 may transmit the encapsulated data in an on-demand manner.
[0062] The receiving device 10004 according to an embodiment includes a receiver 10005, a point cloud video decoder 10006 and / or a renderer 10007. According to an embodiment, the receiving device 10004 may include a device configured to communicate with a base station and / or other wireless devices using a radio access technology (e.g., 5G New RAT (NR), Long Term Evolution (LTE)), a robot, a vehicle, an AR / VR / XR device, a portable device, a home appliance, an Internet of Things (IoT) device, and an AI device / server.
[0063] The receiver 10005 according to the embodiment receives a bitstream containing point cloud video data or a file / segment encapsulated with a bitstream from a network or a storage medium. The receiver 10005 may perform necessary data processing according to a network system (e.g., a communication network system of 4G, 5G, 6G, etc.). The receiver 10005 according to the embodiment may decapsulate the received file / segment and output a bitstream. According to the embodiment, the receiver 10005 may include a decapsulator (or decapsulation module) configured to perform a decapsulation operation. The decapsulator may be implemented as an element (or component) separated from the receiver 10005.
[0064] The point cloud video decoder 10006 decodes the bitstream containing the point cloud video data. The point cloud video decoder 10006 can decode the point cloud video data according to the method in which the point cloud video data is encoded (for example, in the reverse process of the operation of the point cloud video encoder 10002). Therefore, the point cloud video decoder 10006 can decode the point cloud video data by performing point cloud decompression encoding (the reverse process of point cloud compression). The point cloud decompression encoding includes G-PCC encoding.
[0065] The renderer 10007 renders the decoded point cloud video data. The renderer 10007 may output the point cloud content by rendering not only the point cloud video data but also the audio data. According to an embodiment, the renderer 10007 may include a display configured to display the point cloud content. According to an embodiment, the display may be implemented as a separate device or component rather than being included in the renderer 10007.
[0066] The arrow indicated by the dotted line in the figure represents the transmission path of the feedback information obtained by the receiving device 10004. Feedback information is information that reflects the interactivity with the user who consumes the point cloud content, and includes information about the user (e.g., head orientation information, viewport information, etc.). Specifically, when the point cloud content is content for a service that requires interaction with the user (e.g., self-driving service, etc.), the feedback information may be provided to the content sender (e.g., sending device 10000) and / or the service provider. Depending on the embodiment, the feedback information may be used in the receiving device 10004 and the sending device 10000, or may not be provided.
[0067] The head orientation information according to the embodiment is information about the user's head position, orientation, angle, movement, etc. The receiving device 10004 according to the embodiment may calculate the viewport information based on the head orientation information. The viewport information may be information about the area of the point cloud video that the user is watching. The viewpoint is the point through which the user watches the point cloud video, and may refer to the center point of the viewport area. That is, the viewport is an area centered on the viewpoint, and the size and shape of the area may be determined by the field of view (FOV). Therefore, in addition to the head orientation information, the receiving device 10004 may also extract the viewport information based on the vertical or horizontal FOV supported by the device. In addition, the receiving device 10004 performs gaze analysis, etc. to check the way the user consumes the point cloud, the area where the user gazes in the point cloud video, the gaze time, etc. According to the embodiment, the receiving device 10004 may send feedback information including the gaze analysis result to the sending device 10000. The feedback information according to the embodiment may be obtained during the rendering and / or display process. The feedback information according to the embodiment may be obtained by one or more sensors included in the receiving device 10004. According to an embodiment, the feedback information may be obtained by the renderer 10007 or a separate external element (or device, component, etc.). Figure 1The dotted line in represents the process of sending the feedback information obtained by the renderer 10007. The point cloud content providing system can process (encode / decode) the point cloud data based on the feedback information. Therefore, the point cloud video decoder 10006 can perform a decoding operation based on the feedback information. The receiving device 10004 can send the feedback information to the sending device 10000. The sending device 10000 (or the point cloud video data encoder 10002) can perform an encoding operation based on the feedback information. Therefore, the point cloud content providing system can effectively process necessary data (for example, point cloud data corresponding to the user's head position) based on the feedback information instead of processing (encoding / decoding) the entire point cloud data, and provide the point cloud content to the user.
[0068] According to an embodiment, the transmitting device 10000 may be referred to as an encoder, a transmitting device, a transmitter, etc., and the receiving device 10004 may be referred to as a decoder, a receiving device, a receiver, etc.
[0069] According to the implementation mode Figure 1 The point cloud data processed in the point cloud content providing system (through a series of processes of acquisition / encoding / transmission / decoding / rendering) may be referred to as point cloud content data or point cloud video data. According to an embodiment, point cloud content data may be used as a concept covering metadata or signaling information related to point cloud data.
[0070] Figure 1 The elements of the point cloud content providing system shown may be implemented by hardware, software, a processor and / or a combination thereof.
[0071] Figure 2 is a block diagram illustrating point cloud content providing operations according to an embodiment.
[0072] Figure 2 The block diagram shows Figure 1 The operation of the point cloud content providing system described in . As described above, the point cloud content providing system can process point cloud data based on point cloud compression coding (e.g., G-PCC).
[0073] According to an embodiment, a point cloud content providing system (e.g., a point cloud sending device 10000 or a point cloud video acquirer 10001) can acquire a point cloud video (20000). The point cloud video is represented by a point cloud belonging to a coordinate system for expressing a 3D space. According to an embodiment, a point cloud video may include a Ply (Polygon file format or Stanford Triangle format) file. When a point cloud video has one or more frames, the acquired point cloud video may include one or more Ply files. A Ply file contains point cloud data such as point geometry and / or attributes. Geometry includes the position of a point. The position of each point may be represented by a parameter (e.g., the value of the X, Y, and Z axes) representing a three-dimensional coordinate system (e.g., a coordinate system consisting of X, Y, and Z axes). Attributes include attributes of the point (e.g., information about the texture, color (YCbCr or RGB), reflectivity r, transparency, etc. of each point). A point has one or more attributes. For example, a point may have a color attribute or two attributes, color and reflectivity. According to an embodiment, geometry may be referred to as position, geometry information, geometry data, etc., and attributes may be referred to as attributes, attribute information, attribute data, etc. The point cloud content providing system (e.g., the point cloud transmitting device 10000 or the point cloud video acquirer 10001) may obtain point cloud data from information related to a point cloud video acquisition process (e.g., depth information, color information, etc.).
[0074] According to an embodiment, a point cloud content providing system (e.g., a transmitting device 10000 or a point cloud video encoder 10002) may encode point cloud data (20001). The point cloud content providing system may encode point cloud data based on point cloud compression coding. As described above, point cloud data may include geometry and attributes of points. Therefore, the point cloud content providing system may perform geometry coding for encoding geometry and output a geometry bitstream. The point cloud content providing system may perform attribute coding for encoding attributes and output an attribute bitstream. According to an embodiment, the point cloud content providing system may perform attribute coding based on geometry coding. The geometry bitstream and the attribute bitstream according to the embodiment may be multiplexed and output as one bitstream. The bitstream according to the embodiment may also include signaling information related to geometry coding and attribute coding.
[0075] According to the embodiment, the point cloud content providing system (eg, the transmitting device 10000 or the transmitter 10003) may transmit the encoded point cloud data (20002). Figure 1 As shown, the encoded point cloud data may be represented by a geometry bitstream and an attribute bitstream. In addition, the encoded point cloud data may be sent in the form of a bitstream together with signaling information related to encoding of the point cloud data (e.g., signaling information related to geometry encoding and attribute encoding). The point cloud content providing system may encapsulate the bitstream carrying the encoded point cloud data and send it in the form of a file or a fragment.
[0076] The point cloud content providing system (eg, receiving device 10004 or receiver 10005) according to the embodiment may receive a bit stream containing encoded point cloud data. In addition, the point cloud content providing system (eg, receiving device 10004 or receiver 10005) may demultiplex the bit stream.
[0077] The point cloud content providing system (e.g., the receiving device 10004 or the point cloud video decoder 10005) may decode the encoded point cloud data (e.g., geometry bitstream, attribute bitstream) sent in the bitstream. The point cloud content providing system (e.g., the receiving device 10004 or the point cloud video decoder 10005) may decode the point cloud video data based on the signaling information related to the encoding of the point cloud video data contained in the bitstream. The point cloud content providing system (e.g., the receiving device 10004 or the point cloud video decoder 10005) may decode the geometry bitstream to reconstruct the position (geometry) of the point. The point cloud content providing system may reconstruct the attributes of the point by decoding the attribute bitstream based on the reconstructed geometry. The point cloud content providing system (e.g., the receiving device 10004 or the point cloud video decoder 10005) may reconstruct the point cloud video based on the position according to the reconstructed geometry and the decoded attributes.
[0078] According to an embodiment, a point cloud content providing system (e.g., receiving device 10004 or renderer 10007) may render decoded point cloud data (20004). The point cloud content providing system (e.g., receiving device 10004 or renderer 10007) may use various rendering methods to render the geometry and attributes decoded by the decoding process. Points in the point cloud content may be rendered as vertices with a specific thickness, cubes with a specific minimum size centered at the corresponding vertex position, or circles centered at the corresponding vertex position. All or part of the rendered point cloud content is provided to the user through a display (e.g., a VR / AR display, a general display, etc.).
[0079] The point cloud content providing system (e.g., receiving device 10004) according to the embodiment may obtain feedback information (20005). The point cloud content providing system may encode and / or decode the point cloud data based on the feedback information. Feedback Information and Operations of the Point Cloud Content Providing System According to the Embodiment and Reference Figure 1 The feedback information and operations described are the same, so their detailed description is omitted.
[0080] Figure 3 An exemplary process of capturing point cloud video according to an embodiment is shown.
[0081] Figure 3 Show reference Figures 1 to 2 An exemplary point cloud video capture process of a point cloud content providing system is described.
[0082] Point cloud content includes point cloud videos (images and / or videos) representing objects and / or environments located in various 3D spaces (e.g., 3D spaces representing real environments, 3D spaces representing virtual environments, etc.). Therefore, a point cloud content providing system according to an embodiment may use one or more cameras (e.g., an infrared camera capable of obtaining depth information, an RGB camera capable of extracting color information corresponding to the depth information, etc.), a projector (e.g., an infrared pattern projector that obtains depth information), LiDAR, etc. to capture point cloud videos. A point cloud content providing system according to an embodiment may extract a geometric shape composed of points in a 3D space from the depth information and extract the attributes of each point from the color information to obtain point cloud data. Images and / or videos according to an embodiment may be captured based on at least one of an inward-facing technique and an outward-facing technique.
[0083] Figure 3 The left portion of shows inward-facing techniques. Inward-facing techniques refer to techniques that utilize one or more cameras (or camera sensors) positioned around a central object to capture an image of the central object. Inward-facing techniques may be used to generate point cloud content that provides a 360-degree image of a key object to a user (e.g., VR / AR content that provides a 360-degree image of an object (e.g., a key object such as a character, player, object, or actor) to a user).
[0084] Figure 3 The right portion of shows outward-facing techniques. Outward-facing techniques refer to techniques that utilize one or more cameras (or camera sensors) positioned around a central object to capture the environment of the central object rather than an image of the central object. Outward-facing techniques can be used to generate point cloud content that provides the surrounding environment as it appears from the user's perspective (e.g., content representing the external environment that can be provided to a user of a self-driving vehicle).
[0085] As shown in the figure, the point cloud content may be generated based on the capture operation of one or more cameras. In this case, the coordinate system may be different between the cameras, so the point cloud content providing system may calibrate one or more cameras to set the global coordinate system before the capture operation. In addition, the point cloud content providing system may generate point cloud content by synthesizing any image and / or video with the image and / or video captured by the above-mentioned capture technology. The point cloud content providing system may not perform when generating point cloud content representing the virtual space. Figure 3 The point cloud content providing system according to the embodiment may perform post-processing on the captured image and / or video. In other words, the point cloud content providing system may remove unwanted areas (e.g., background), identify the space to which the captured image and / or video is connected, and when there is a space hole, perform an operation to fill the space hole.
[0086] The point cloud content providing system can generate a point cloud content by performing coordinate transformation on the points of the point cloud video obtained from each camera. The point cloud content providing system can perform coordinate transformation on the points based on the position coordinates of each camera. Therefore, the point cloud content providing system can generate content representing a wide range, or can generate point cloud content with high density points.
[0087] Figure 4 An exemplary point cloud encoder according to an implementation is shown.
[0088] Figure 4 Show Figure 1 An example of a point cloud video encoder 10002. The point cloud encoder reconstructs and encodes point cloud data (e.g., the location and / or attributes of a point) to adjust the quality of the point cloud content (e.g., lossless, lossy, or near-lossless) according to network conditions or applications. When the total size of the point cloud content is large (e.g., 60 Gbps of point cloud content given for 30 fps), the point cloud content providing system may not be able to stream the content in real time. Therefore, the point cloud content providing system may reconstruct the point cloud content based on the maximum target bit rate to provide the point cloud content according to the network environment, etc.
[0089] As reference Figures 1 to 2 As described above, the point cloud encoder can perform both geometry encoding and attribute encoding. Geometry encoding is performed before attribute encoding.
[0090] According to an embodiment, the point cloud encoder includes a coordinate transformer (transforming coordinates) 40000, a quantizer (quantizing and removing points (voxelization)) 40001, an octree analyzer (analyzing octree) 40002 and a surface approximation analyzer (analyzing surface approximation) 40003, an arithmetic encoder (arithmetic coding) 40004, a geometry reconstructor (reconstructing geometry) 40005, a color transformer (transforming color) 40006, an attribute transformer (transforming attributes) 40007, a RAHT transformer (RAHT) 40008, an LOD generator (generating LOD) 40009, a lifting transformer (lifting) 40010, a coefficient quantizer (quantizing coefficients) 40011 and / or an arithmetic encoder (arithmetic coding) 40012.
[0091] The coordinate transformer 40000, the quantizer 40001, the octree analyzer 40002, the surface approximation analyzer 40003, the arithmetic encoder 40004 and the geometry reconstructor 40005 may perform geometry coding. The geometry coding according to the embodiment may include octree geometry coding, direct coding, triplet geometry coding and entropy coding. Direct coding and triplet geometry coding are applied selectively or in combination. The geometry coding is not limited to the above examples.
[0092] As shown in the figure, the coordinate transformer 40000 according to the embodiment receives the position and transforms it into coordinates. For example, the position can be transformed into position information in a three-dimensional space (e.g., a three-dimensional space represented by an XYZ coordinate system). The position information in the three-dimensional space according to the embodiment can be referred to as geometric information.
[0093] The quantizer 40001 according to the embodiment quantizes the geometry. For example, the quantizer 40001 may quantize the points based on the minimum position value of all points (e.g., the minimum value on each of the X, Y, and Z axes). The quantizer 40001 performs a quantization operation: the difference between the minimum position value and the position value of each point is multiplied by a preset quantization scale value, and then the nearest integer value is found by rounding the value obtained by the multiplication. Therefore, one or more points may have the same quantized position (or position value). The quantizer 40001 according to the embodiment performs voxelization based on the quantized position to reconstruct the quantized point. As in the case of a pixel (the smallest unit containing 2D image / video information), a point of a point cloud content (or 3D point cloud video) according to the embodiment may be included in one or more voxels. As a composite of volume and pixel, the term voxel refers to a 3D cubic space generated when a 3D space is divided into units (unit = 1.0) based on axes representing the 3D space (e.g., X-axis, Y-axis, and Z-axis). The quantizer 40001 may match a point group in a 3D space with a voxel. According to an embodiment, a voxel may include only one point. According to an embodiment, a voxel may include one or more points. In order to represent a voxel as a point, the position of the center of the voxel may be set based on the position of one or more points included in the voxel. In this case, the properties of all positions included in a voxel may be combined and assigned to the voxel.
[0094] The octree analyzer 40002 according to the embodiment performs octree geometry encoding (or octree encoding) to present voxels in an octree structure. The octree structure represents points matched to voxels based on an octal tree structure.
[0095] The surface approximation analyzer 40003 according to an embodiment may analyze and approximate an octree. The octree analysis and approximation according to an embodiment is a process of analyzing a region including a plurality of points to efficiently provide an octree and voxelization.
[0096] The arithmetic encoder 40004 according to the embodiment performs entropy coding on the octree and / or the approximate octree. For example, the coding scheme includes arithmetic coding. As a result of the coding, a geometry bitstream is generated.
[0097] The color converter 40006, the attribute converter 40007, the RAHT converter 40008, the LOD generator 40009, the lifting converter 40010, the coefficient quantizer 40011 and / or the arithmetic encoder 40012 perform attribute encoding. As described above, a point may have one or more attributes. The attribute encoding according to the embodiment is also applied to the attributes of a point. However, when the attribute (e.g., color) includes one or more elements, the attribute encoding is applied independently to each element. The attribute encoding according to the embodiment includes color transform encoding, attribute transform encoding, regional adaptive hierarchical transform (RAHT) encoding, interpolation-based hierarchical nearest neighbor prediction (prediction transform) encoding, and interpolation-based hierarchical nearest neighbor prediction (lifting transform) encoding with an update / lifting step. Depending on the point cloud content, the above-mentioned RAHT encoding, prediction transform encoding, and lifting transform encoding may be selectively used, or a combination of one or more encoding schemes may be used. The attribute encoding according to the embodiment is not limited to the above examples.
[0098] The color converter 40006 according to the embodiment performs color conversion encoding to convert the color value (or texture) included in the attribute. For example, the color converter 40006 can convert the format of the color information (e.g., from RGB to YCbCr). Optionally, the operation of the color converter 40006 according to the embodiment can be applied according to the color value included in the attribute.
[0099] The geometry reconstructor 40005 according to an embodiment reconstructs (decompresses) an octree and / or an approximate octree. The geometry reconstructor 40005 reconstructs an octree / voxel based on the result of analyzing the point distribution. The reconstructed octree / voxel may be referred to as a reconstructed geometry (restored geometry).
[0100] The attribute transformer 40007 according to the embodiment performs attribute transformation to transform the attribute based on the reconstructed geometry and / or the position where the geometry encoding is not performed. As described above, since the attribute depends on the geometry, the attribute transformer 40007 can transform the attribute based on the reconstructed geometry information. For example, based on the position value of the point included in the voxel, the attribute transformer 40007 can transform the attribute of the point at the position. As described above, when the center position of the voxel is set based on the position of one or more points included in the voxel, the attribute transformer 40007 transforms the attribute of one or more points. When triplet geometry encoding is performed, the attribute transformer 40007 can transform the attribute based on the triplet geometry encoding.
[0101] The attribute converter 40007 can perform attribute conversion by calculating the average of the attributes or attribute values (e.g., the color or reflectivity of each point) of the neighboring points within a specific position / radius from the center position (or position value) of each voxel. The attribute converter 40007 can apply weights according to the distance from the center to each point when calculating the average. Therefore, each voxel has a position and a calculated attribute (or attribute value).
[0102] The attribute converter 40007 can search for neighboring points within a specific position / radius from the center position of each voxel based on a KD tree or a Morton code. The KD tree is a binary search tree and supports a data structure capable of managing points based on position, so that a nearest neighbor search (NNS) can be performed quickly. The Morton code is generated by presenting the coordinates (e.g., (x, y, z)) representing the 3D positions of all points as bit values and mixing the bits. For example, when the coordinates representing the point position are (5, 9, 1), the bit values of the coordinates are (0101, 1001, 0001). Mixing the bit values in the order of z, y, and x according to the bit index produces 010001000111. This value is represented as a decimal number 1095. That is, the Morton code value of the point with coordinates (5, 9, 1) is 1095. The attribute converter 40007 can sort the points based on the Morton code value and perform NNS through a depth-first traversal process. After the attribute transform operation, KD tree or Morton code is used when NNS is needed in another transform process for attribute encoding.
[0103] As shown, the transformed attributes are input to the RAHT transformer 40008 and / or the LOD generator 40009.
[0104] The RAHT transformer 40008 according to an embodiment performs RAHT encoding for predicting attribute information based on the reconstructed geometric information. For example, the RAHT transformer 40008 may predict attribute information of a higher-level node in the octree based on attribute information associated with a lower-level node in the octree.
[0105] The LOD generator 40009 according to an embodiment generates a level of detail (LOD) to perform predictive transform coding. The LOD according to an embodiment is the degree of detail of the point cloud content. As the LOD value decreases, the details of the point cloud content are degraded. As the LOD value increases, the details of the point cloud content are enhanced. Points can be classified by LOD.
[0106] The lifting transformer 40010 according to an embodiment performs lifting transform coding that transforms point cloud attributes based on weights. As described above, lifting transform coding can be optionally applied.
[0107] The coefficient quantizer 40011 according to the embodiment quantizes the attribute of the attribute encoding based on the coefficient.
[0108] The arithmetic encoder 40012 according to the embodiment encodes quantized properties based on arithmetic coding.
[0109] Although not shown in the figure, Figure 4 The elements of the point cloud encoder may be implemented by hardware, software, firmware, or a combination thereof including one or more processors or integrated circuits configured to communicate with one or more memories included in the point cloud providing device. The one or more processors may perform the above Figure 4 At least one of the operations and / or functions of the elements of the point cloud encoder. In addition, one or more processors may be operable or executable to perform Figure 4 The software program and / or instruction set for the operation and / or function of the elements of the point cloud encoder. According to the embodiment, one or more memories may include high-speed random access memory, or include non-volatile memory (e.g., one or more disk storage devices, flash memory devices or other non-volatile solid-state memory devices).
[0110] Figure 5 An example of a voxel according to an embodiment is shown.
[0111] Figure 5 The voxels are shown in a 3D space represented by a coordinate system consisting of three axes (X axis, Y axis and Z axis). Figure 4 As described above, the point cloud encoder (eg, quantizer 40001) may perform voxelization. A voxel refers to a 3D cubic space generated when a 3D space is divided into units (unit=1.0) based on axes representing the 3D space (eg, X-axis, Y-axis, and Z-axis). Figure 5 An example of voxels generated by an octree structure is shown, where a cubic axis-aligned bounding box defined by two poles (0,0,0) and (2d,2d,2d) is recursively subdivided. A voxel includes at least one point. The spatial coordinates of the voxel can be estimated from the positional relationship with the voxel group. As described above, the voxel has properties similar to the pixels of a 2D image / video (e.g., color or reflectivity). The details of the voxel are similar to those of the reference Figure 4 Those described are the same, so their description is omitted.
[0112] Figure 6 An example of an octree and occupancy code according to an embodiment is shown.
[0113] As reference Figures 1 to 4 As described, a point cloud content providing system (point cloud video encoder 10002) or a point cloud encoder (e.g., octree analyzer 40002) performs octree geometry encoding (or octree encoding) based on an octree structure to efficiently manage the area and / or position of voxels.
[0114] Figure 6 The upper part of FIG. 1 shows the octree structure. The 3D space of the point cloud content according to the embodiment is represented by the axes (eg, X-axis, Y-axis, and Z-axis) of the coordinate system. d ,2 d ,2 d ) to create an octree structure. Here, 2 d It can be set to the value of the minimum bounding box that surrounds all points of the point cloud content (or point cloud video). Here, d represents the depth of the octree. The value of d is determined in the following formula. In the following formula, (x int n ,y int n ,z int n ) represents the position (or position value) of the quantized point.
[0115]
[0116] like Figure 6 As shown in the middle of the upper part of , the entire 3D space can be divided into eight spaces according to the partition. Each divided space is represented by a cube with six faces. Figure 6 As shown in the upper right part of , each of the eight spaces is divided again based on the axis of the coordinate system (e.g., X-axis, Y-axis, and Z-axis). Therefore, each space is divided into eight smaller spaces. The divided smaller spaces are also represented by cubes with six faces. This partitioning scheme is applied until the leaf nodes of the octree become voxels.
[0117] Figure 6 The lower part shows the octree occupancy code. The occupancy code of the octree is generated to indicate whether each of the eight divided spaces generated by dividing one space contains at least one point. Therefore, a single occupancy code is represented by eight child nodes. Each child node represents the occupancy of the divided space, and the child node has a 1-bit value. Therefore, the occupancy code is represented as an 8-bit code. That is, when the space corresponding to the child node contains at least one point, the node is assigned a value of 1. When the space corresponding to the child node does not contain a point (the space is empty), the node is assigned a value of 0. Since Figure 6The occupancy code shown is 00100001, so it indicates that the spaces corresponding to the third and eighth subnodes among the eight subnodes each contain at least one point. As shown in the figure, each of the third and eighth subnodes has eight subnodes, and the subnodes are represented by 8-bit occupancy codes. The accompanying drawings show that the occupancy code of the third subnode is 10000111, and the occupancy code of the eighth subnode is 01001111. The point cloud encoder (e.g., arithmetic encoder 40004) according to the embodiment may perform entropy encoding on the occupancy code. In order to increase the compression efficiency, the point cloud encoder may perform intra-frame / inter-frame encoding on the occupancy code. The receiving device (e.g., receiving device 10004 or point cloud video decoder 10006) according to the embodiment reconstructs the occupancy code based on the occupancy code.
[0118] According to an embodiment, a point cloud encoder (e.g., Figure 4 The point cloud encoder or octree analyzer 40002) may perform voxelization and octree encoding to store point locations. However, points are not always evenly distributed in the 3D space, so there may be specific areas where fewer points exist. Therefore, performing voxelization on the entire 3D space is inefficient. For example, when a specific area contains very few points, voxelization does not need to be performed in the specific area.
[0119] Therefore, for the above-mentioned specific area (or nodes other than the leaf nodes of the octree), the point cloud encoder according to the embodiment may skip voxelization and perform direct encoding to directly encode the point positions included in the specific area. The coordinates of the directly encoded points according to the embodiment are called direct coding mode (DCM). The point cloud encoder according to the embodiment may also perform triplet geometry encoding based on the surface model, which is to reconstruct the point positions in the specific area (or node) based on voxels. Triplet geometry encoding is a geometry encoding that represents an object as a series of triangular meshes. Therefore, the point cloud decoder can generate a point cloud from the mesh surface. Direct encoding and triplet geometry encoding according to the embodiment can be performed selectively. In addition, direct encoding and triplet geometry encoding according to the embodiment can be performed in combination with octree geometry encoding (or octree encoding).
[0120] In order to perform direct encoding, the option of using direct mode to apply direct encoding should be enabled. The node to which direct encoding is to be applied is not a leaf node, and there should be points less than a threshold value within a specific node. In addition, the total number of points to which direct encoding is to be applied should not exceed a preset threshold. When the above conditions are met, the point cloud encoder (or arithmetic encoder 40004) according to the embodiment can perform entropy encoding on the point position (or position value).
[0121] A point cloud encoder (e.g., surface approximation analyzer 40003) according to an embodiment may determine a specific level of the octree (a level less than the depth d of the octree), and may use a surface model starting from this level to perform triplet geometry encoding to reconstruct point positions in the node area based on voxels (triplet mode). A point cloud encoder according to an embodiment may specify a level to which triplet geometry encoding is to be applied. For example, when a specific level is equal to the depth of the octree, the point cloud encoder does not operate in triplet mode. In other words, a point cloud encoder according to an embodiment may operate in triplet mode only when the specified level is less than the depth value of the octree. A 3D cubic area of a node at a specified level according to an embodiment is referred to as a block. A block may include one or more voxels. A block or voxel may correspond to a block. Geometry is represented as a surface within each block. A surface according to an embodiment may intersect each edge of a block at most once.
[0122] A block has 12 edges, so there are at least 12 intersections in a block. Each intersection is called a vertex. When there is at least one occupied voxel adjacent to the edge among all blocks sharing the edge, the vertex existing along the edge is detected. An occupied voxel according to an embodiment refers to a voxel containing a point. The vertex position detected along the edge is the average position of the edge of all voxels adjacent to the edge among all blocks along the shared edge.
[0123] Once the vertex is detected, the point cloud encoder according to the embodiment can perform entropy encoding on the start point (x, y, z) of the edge, the direction vector (Δx, Δy, Δz) of the edge, and the vertex position value (relative position value within the edge). When triplet geometry encoding is applied, the point cloud encoder according to the embodiment (e.g., geometry reconstructor 40005) can generate restored geometry (reconstructed geometry) by performing triangle reconstruction, upsampling, and voxelization processes.
[0124] The vertices located at the edges of the block determine the surface passing through the block. The surface according to an embodiment is a non-planar polygon. In the triangle reconstruction process, the surface represented by the triangle is reconstructed based on the starting point of the edge, the direction vector of the edge and the position value of the vertex. The triangle reconstruction process is performed by i) calculating the centroid value of each vertex, ii) subtracting the center value from each vertex value, and iii) estimating the sum of the squares of the values obtained by subtraction.
[0125] ① ② ③
[0126] The minimum value of the sum is estimated, and the projection process is performed according to the axis with the minimum value. For example, when the element x is the smallest, each vertex is projected on the x-axis relative to the center of the block and projected on the (y,z) plane. When the value obtained by the projection on the (y,z) plane is (ai,bi), the value of θ is estimated by atan2(bi,ai), and the vertices are sorted based on the value of θ. The following table shows that vertex combinations for creating triangles according to the number of vertices. The vertices are sorted from 1 to n. The following table shows that for four vertices, two triangles can be constructed according to the vertex combination. The first triangle can be composed of vertices 1, 2, and 3 among the sorted vertices, and the second triangle can be composed of vertices 3, 4, and 1 among the sorted vertices.
[0127] Table 2-1. Triangles formed from vertices ordered 1,…,n
[0128] n triangle
[0129] 3(1,2,3)
[0130] 4(1,2,3),(3,4,1)
[0131] 5(1,2,3),(3,4,5),(5,1,3)
[0132] 6(1,2,3),(3,4,5),(5,6,1),(1,3,5)
[0133] 7(1,2,3),(3,4,5),(5,6,7),(7,1,3),(3,5,7)
[0134] 8(1,2,3),(3,4,5),(5,6,7),(7,8,1),(1,3,5),(5,7,1)
[0135] 9(1,2,3),(3,4,5),(5,6,7),(7,8,9),(9,1,3),(3,5,7),(7,9,3)
[0136] 10(1,2,3),(3,4,5),(5,6,7),(7,8,9),(9,10,1),(1,3,5),(5,7,9),(9,1,5)
[0137] 11(1,2,3),(3,4,5),(5,6,7),(7,8,9),(9,10,11),(11,1,3),(3,5,7),(7,9,11),(11,3,7)
[0138] 12(1,2,3),(3,4,5),(5,6,7),(7,8,9),(9,10,11),(11,12,1),(1,3,5),(5,7,9),(9,11,1),(1,5,9)
[0139] An upsampling process is performed to add points in the middle along the edges of the triangle, and voxelization is performed. The added points are generated based on the upsampling factor and the width of the block. The added points are called refined vertices. The point cloud encoder according to an embodiment may voxelize the refined vertices. In addition, the point cloud encoder may perform attribute encoding based on the voxelized position (or position value).
[0140] Figure 7 An example of a neighbor node pattern according to an embodiment is shown.
[0141] In order to increase the compression efficiency of the point cloud video, the point cloud encoder according to an embodiment may perform entropy coding based on context adaptive arithmetic coding.
[0142] As reference Figures 1 to 6 As described, the point cloud content providing system or point cloud encoder (eg, point cloud video encoder 10002, Figure 4 The point cloud encoder or arithmetic encoder 40004) may immediately perform entropy encoding on the occupancy code. In addition, the point cloud content providing system or the point cloud encoder may perform entropy encoding (intra-frame encoding) based on the occupancy code of the current node and the occupancy of the neighboring nodes, or perform entropy encoding (inter-frame encoding) based on the occupancy code of the previous frame. The frame representation according to the embodiment is a collection of point cloud videos generated simultaneously. The compression efficiency of the intra-frame coding / inter-frame coding according to the embodiment may depend on the number of neighboring nodes referenced. When the number of bits increases, the operation becomes complicated, but the encoding can be biased to one side, which can increase the compression efficiency. For example, when a 3-bit context is given, 2 bits need to be used. 3 = 8 ways to perform encoding. The division of parts for encoding affects the implementation complexity. Therefore, it is necessary to meet appropriate levels of compression efficiency and complexity.
[0143] Figure 7 The process of obtaining an occupancy pattern based on the occupancy of neighbor nodes is shown. The point cloud encoder according to the embodiment determines the occupancy of neighbor nodes of each node of the octree and obtains the value of the neighbor pattern. The neighbor node pattern is used to infer the occupancy pattern of the node. Figure 7 The upper part of the diagram shows a cube corresponding to a node (the cube in the middle) and six cubes (neighboring nodes) that share at least one face with the cube. The nodes shown in the diagram are nodes at the same depth. The numbers shown in the diagram represent the weights (1, 2, 4, 8, 16, and 32) associated with the six nodes, respectively. The weights are assigned in sequence according to the positions of the neighboring nodes.
[0144] Figure 7 The lower part of shows the neighbor node pattern value. The neighbor node pattern value is the sum of the values multiplied by the weights of the occupied neighbor nodes (neighbor nodes with points). Therefore, the neighbor node pattern value is 0 to 63. When the neighbor node pattern value is 0, it indicates that there are no nodes with points among the neighbor nodes of the node (no occupied nodes). When the neighbor node pattern value is 63, it indicates that all neighbor nodes are occupied nodes. As shown in the figure, since the neighbor nodes assigned weights 1, 2, 4 and 8 are occupied nodes, the neighbor node pattern value is 15 (the sum of 1, 2, 4 and 8). The point cloud encoder can perform encoding according to the neighbor node pattern value (for example, when the neighbor node pattern value is 63, 64 types of encoding can be performed). According to an embodiment, the point cloud encoder can reduce the encoding complexity by changing the neighbor node pattern value (for example, based on a table that changes 64 to 10 or 6).
[0145] Figure 8 An example of point configuration in each LOD according to an embodiment is shown.
[0146] As reference Figures 1 to 7 As described, the encoded geometry is reconstructed (decompressed) before performing attribute encoding. When direct encoding is applied, the geometry reconstruction operation may include changing the placement of the directly encoded points (e.g., placing the directly encoded points in front of the point cloud data). When triplet geometry encoding is applied, the geometry reconstruction process is performed by triangle reconstruction, upsampling, and voxelization. Since the attributes depend on the geometry, attribute encoding is performed based on the reconstructed geometry.
[0147] The point cloud encoder (e.g., LOD generator 40009) can classify (reorganize) the points by LOD. The figure shows the point cloud content corresponding to the LOD. The leftmost screen in the figure shows the original point cloud content. The second screen from the left in the figure shows the distribution of points in the lowest LOD, and the rightmost screen in the figure shows the distribution of points in the highest LOD. That is, the points in the lowest LOD are sparsely distributed, and the points in the highest LOD are densely distributed. That is, as the LOD increases in the direction indicated by the arrow indicated at the bottom of the figure, the space (or distance) between the points becomes narrower.
[0148] Figure 9 An example of point configuration for each LOD according to an embodiment is shown.
[0149] As reference Figures 1 to 8 As described, the point cloud content providing system or point cloud encoder (eg, point cloud video encoder 10002, Figure 4The point cloud encoder or LOD generator 40009) can generate LOD. LOD is generated by reorganizing points into a set of refinement levels according to the set LOD distance value (or Euclidean distance set). The LOD generation process is performed not only by the point cloud encoder, but also by the point cloud decoder.
[0150] Figure 9 The upper part of shows examples of points (P0 to P9) of point cloud content distributed in 3D space. Figure 9 In , the original order represents the order of points P0 to P9 before LOD generation. Figure 9 In the LOD-based order, the order of points generated according to the LOD is shown. Points are reorganized by LOD. In addition, a high LOD contains points belonging to a lower LOD. Figure 9 As shown, LOD0 contains P0, P5, P4, and P2. LOD1 contains the points of LOD0, P1, P6, and P3. LOD2 contains the points of LOD0, the points of LOD1, P9, P8, and P7.
[0151] As reference Figure 4 As described, the point cloud encoder according to the embodiment may selectively or in combination perform prediction transform coding, lifting transform coding, and RAHT transform coding.
[0152] The point cloud encoder according to the embodiment may generate a predictor for a point to perform predictive transform coding for setting a prediction attribute (or a prediction attribute value) of each point. That is, N predictors may be generated for N points. The predictor according to the embodiment may calculate a weight (=1 / distance) based on the LOD value of each point, index information about neighboring points that exist within a set distance of each LOD, and the distance to the neighboring point.
[0153] The predicted attribute (or attribute value) according to the embodiment is set to the average of the values obtained by multiplying the attribute (or attribute value) (e.g., color, reflectivity, etc.) of the neighboring points set in the predictor of each point by the weight (or weight value) calculated based on the distance to each neighboring point. The point cloud encoder (e.g., coefficient quantizer 40011) according to the embodiment may quantize and inversely quantize the residual (which may be referred to as residual attribute, residual attribute value, or attribute prediction residual) obtained by subtracting the predicted attribute (attribute value) from the attribute (attribute value) of each point. The quantization process is configured as shown in the following table.
[0154] Table. Attribute prediction residual quantization pseudo code
[0155] int PCCQuantization(int value,int quantStep){
[0156] if(value>=0){
[0157] return floor(value / quantStep+1.0 / 3.0);
[0158] }else{
[0159] return-floor(-value / quantStep+1.0 / 3.0);
[0160] }
[0161] }
[0162] Table. Pseudocode for inverse quantization of attribute prediction residuals
[0163] int PCCInverseQuantization(int value,int quantStep){
[0164] if(quantStep==0){
[0165] return value;
[0166] }else{
[0167] return value*quantStep;
[0168] }
[0169] }
[0170] When the predictor of each point has a neighboring point, the point cloud encoder (e.g., the arithmetic encoder 40012) according to the embodiment may perform entropy encoding on the quantized and inverse quantized residual values as described above. When the predictor of each point has no neighboring point, the point cloud encoder (e.g., the arithmetic encoder 40012) according to the embodiment may perform entropy encoding on the attributes of the corresponding point without performing the above operation.
[0171] The point cloud encoder (e.g., lifting transformer 40010) according to an embodiment may generate a predictor for each point, set the calculated LOD and register neighbor points in the predictor, and set weights according to the distance to the neighbor points to perform lifting transform encoding. The lifting transform encoding according to an embodiment is similar to the above-mentioned predictive transform encoding, but is different in that weights are cumulatively applied to attribute values. The process of cumulatively applying weights to attribute values according to an embodiment is configured as follows.
[0172] 1) Create an array quantized weight (QW) for storing weight values of each point. The initial value of all elements of QW is 1.0. Multiply the QW value of the predictor index of the neighbor node registered in the predictor by the weight of the predictor of the current point, and add the values obtained by multiplication.
[0173] 2) Boosting prediction process: The value obtained by multiplying the attribute value of the point by the weight is subtracted from the existing attribute value to calculate the predicted attribute value.
[0174] 3) Create temporary arrays called updateweight and update, and initialize the temporary arrays to zero.
[0175] 4) Accumulate the weight calculated by multiplying the weight calculated for all predictors by the weight stored in the QW corresponding to the predictor index with the updateweight array as the index of the neighbor node. Accumulate the value obtained by multiplying the attribute value of the neighbor node index by the calculated weight with the update array.
[0176] 5) Boosting update process: Divide the attribute values of the update array of all predictors by the weight value of the updateweight array of the predictor index, and add the existing attribute value to the value obtained by the division.
[0177] 6) For all predictors, the predicted attribute is calculated by multiplying the attribute value updated by the lifting update process by the weight updated by the lifting prediction process (stored in QW). The point cloud encoder (e.g., coefficient quantizer 40011) according to the embodiment quantizes the predicted attribute value. In addition, the point cloud encoder (e.g., arithmetic encoder 40012) performs entropy encoding on the quantized attribute value.
[0178] A point cloud encoder according to an embodiment (e.g., RAHT transformer 40008) may perform RAHT transform coding, in which attributes associated with nodes at a lower level in an octree are used to predict attributes of nodes at a higher level. RAHT transform coding is an example of attribute intra-frame coding by scanning backward through an octree. A point cloud encoder according to an embodiment scans the entire area starting from voxels and repeats a merging process of merging voxels into larger blocks at each step until the root node is reached. The merging process according to an embodiment is performed only on occupied nodes. The merging process is not performed on empty nodes. The merging process is performed on the upper node directly above the empty node.
[0179] The following formula represents the RAHT transformation matrix. In this formula, Represents the average attribute value of voxels at level l. Can be based on and to calculate. and The weight is and
[0180]
[0181] here, Is the low-pass value and is used during the next highest level of merging. Denotes the high-pass coefficient. The high-pass coefficient of each step is quantized and subjected to entropy coding (eg, encoded by arithmetic encoder 400012). The weight is calculated as pass and Create the root node as follows.
[0182]
[0183] Figure 10 A point cloud decoder according to an embodiment is shown.
[0184] Figure 10 The point cloud decoder shown is Figure 1 An example of a point cloud video decoder 10006 described in Figure 1 The operations of the point cloud video decoder 10006 shown are the same or similar operations. As shown, the point cloud decoder can receive a geometry bitstream and an attribute bitstream contained in one or more bitstreams. The point cloud decoder includes a geometry decoder and an attribute decoder. The geometry decoder performs geometry decoding on the geometry bitstream and outputs the decoded geometry. The attribute decoder performs attribute decoding based on the decoded geometry and attribute bitstreams, and outputs the decoded attributes. The decoded geometry and the decoded attributes are used to reconstruct the point cloud content (decoded point cloud).
[0185] Figure 11 A point cloud decoder according to an embodiment is shown.
[0186] Figure 11 The point cloud decoder shown is Figure 10 An example of a point cloud decoder is shown, and a decoding operation can be performed, which is Figures 1 to 9 The inverse process of the encoding operation of the point cloud encoder shown.
[0187] As reference Figure 1 and Figure 10 As described, the point cloud decoder can perform geometry decoding and attribute decoding. Geometry decoding is performed before attribute decoding.
[0188] According to an embodiment, a point cloud decoder includes an arithmetic decoder (arithmetic decoding) 11000, an octree synthesizer (synthesized octree) 11001, a surface approximation synthesizer (synthesized surface approximation) 11002 and a geometry reconstructor (reconstructed geometry) 11003, an inverse coordinate transformer (inverse transformed coordinates) 11004, an arithmetic decoder (arithmetic decoding) 11005, an inverse quantizer (inverse quantization) 11006, a RAHT transformer 11007, an LOD generator (generate LOD) 11008, an inverse lifting (inverse lifting) 11009 and / or an inverse color transformer (inverse transformed color) 11010.
[0189] The arithmetic decoder 11000, the octree synthesizer 11001, the surface approximation synthesizer 11002, the geometry reconstructor 11003, and the coordinate inverse transformer 11004 may perform geometry decoding. The geometry decoding according to the embodiment may include direct encoding and triplet geometry decoding. Direct encoding and triplet geometry decoding are selectively applied. The geometry decoding is not limited to the above examples, and as a reference Figures 1 to 9 The inverse process of the geometric encoding described is performed.
[0190] The arithmetic decoder 11000 according to the embodiment decodes the received geometry bitstream based on arithmetic coding. The operation of the arithmetic decoder 11000 corresponds to the inverse process of the arithmetic encoder 40004.
[0191] The octree synthesizer 11001 according to the embodiment may generate an octree by acquiring an occupancy code (or information about the geometry obtained as a result of decoding) from the decoded geometry bitstream. Figures 1 to 9 Describe the configuration in detail.
[0192] When triplet geometry encoding is applied, a surface approximation synthesizer 11002 according to an embodiment may synthesize a surface based on the decoded geometry and / or the generated octree.
[0193] According to an embodiment, the geometry reconstructor 11003 can regenerate the geometry based on the surface and / or the decoded geometry. Figures 1 to 9 As described, direct coding and triplet geometry coding are selectively applied. Therefore, the geometry reconstructor 11003 directly imports the position information about the points to which direct coding is applied and adds them. When triplet geometry coding is applied, the geometry reconstructor 11003 can reconstruct the geometry by performing the reconstruction operations (e.g., triangle reconstruction, upsampling, and voxelization) of the geometry reconstructor 40005. Details and References Figure 6 Those described are the same, so their description is omitted. The reconstructed geometry may include a point cloud screen or frame that does not contain attributes.
[0194] The coordinate inverse transformer 11004 according to an embodiment may acquire a point position by transforming the coordinates based on the reconstructed geometry.
[0195] The arithmetic decoder 11005, the inverse quantizer 11006, the RAHT transformer 11007, the LOD generator 11008, the inverse lifter 11009 and / or the color inverse transformer 11010 may perform reference Figure 10 Attribute decoding described. Attribute decoding according to an embodiment includes regional adaptive hierarchical transform (RAHT) decoding, hierarchical nearest neighbor prediction (prediction transform) decoding based on interpolation, and hierarchical nearest neighbor prediction (lifting transform) decoding based on interpolation with an update / lifting step. The above three decoding schemes can be used selectively, or a combination of one or more decoding schemes can be used. Attribute decoding according to an embodiment is not limited to the above examples.
[0196] The arithmetic decoder 11005 according to the embodiment decodes the attribute bit stream through arithmetic coding.
[0197] The inverse quantizer 11006 according to the embodiment inversely quantizes information about the decoded attribute bitstream or the attribute obtained as a result of decoding, and outputs the inversely quantized attribute (or attribute value). The inverse quantization may be selectively applied based on the attribute encoding of the point cloud encoder.
[0198] According to an embodiment, the RAHT transformer 11007, the LOD generator 11008 and / or the inverse lifter 11009 may process the reconstructed geometry and inverse quantized attributes. As described above, the RAHT transformer 11007, the LOD generator 11008 and / or the inverse lifter 11009 may selectively perform a decoding operation corresponding to the encoding of the point cloud encoder.
[0199] The color inverse transformer 11010 according to an embodiment performs inverse transform encoding to inversely transform a color value (or texture) included in the decoded attribute. The operation of the color inverse transformer 11010 may be selectively performed based on the operation of the color transformer 40006 of the point cloud encoder.
[0200] Although not shown in the figure, Figure 11 The elements of the point cloud decoder may be implemented by hardware, software, firmware, or a combination thereof including one or more processors or integrated circuits configured to communicate with one or more memories included in the point cloud providing device. The one or more processors may execute the above Figure 11 At least one or more of the operations and / or functions of the elements of the point cloud decoder. In addition, one or more processors may be operable or executed to perform Figure 11 A software program and / or instruction set for the operation and / or functionality of the elements of a point cloud decoder.
[0201] Figure 12 A transmitting device according to an embodiment is shown.
[0202] Figure 12 The sending device shown is Figure 1 The sending device 10000 (or Figure 4 Example of a point cloud encoder). Figure 12 The sending device shown can be executed with reference to Figures 1 to 9 The transmitting device according to the embodiment may include a data input unit 12000, a quantization processor 12001, a voxelization processor 12002, an octree occupancy code generator 12003, a surface model processor 12004, an intra-frame / inter-frame coding processor 12005, an arithmetic encoder 12006, a metadata processor 12007, a color transformation processor 12008, an attribute transformation processor 12009, a prediction / lifting / RAHT transformation processor 12010, an arithmetic encoder 12011 and / or a transmission processor 12012.
[0203] The data input unit 12000 according to the embodiment receives or acquires point cloud data. The data input unit 12000 may perform the same operation and / or acquisition method as the point cloud video acquirer 10001 (or refer to Figure 2 The acquisition process 20000) described above has the same or similar operations and / or acquisition methods.
[0204] The data input unit 12000, the quantization processor 12001, the voxelization processor 12002, the octree occupancy code generator 12003, the surface model processor 12004, the intra / inter coding processor 12005 and the arithmetic encoder 12006 perform geometric coding. Figures 1 to 9 The geometrical encodings described are the same or similar, so their detailed description is omitted.
[0205] According to an embodiment, the quantization processor 12001 quantizes geometry (e.g., position values of points). The operation and / or quantization of the quantization processor 12001 is similar to the reference Figure 4 The operation and / or quantization of the quantizer 40001 described above are the same or similar. Figures 1 to 9 Same as those described.
[0206] According to the embodiment, the voxelization processor 12002 voxelizes the quantized position value of the point. The voxelization processor 120002 may perform the same as the reference Figure 4 The operation and / or voxelization process of the quantizer 40001 described above is the same or similar to the operation and / or process described above. Figures 1 to 9 Same as those described.
[0207] The octree occupancy code generator 12003 according to the embodiment performs octree encoding on the voxelized position of the point based on the octree structure. The octree occupancy code generator 12003 may generate an occupancy code. The octree occupancy code generator 12003 may perform the same as the reference code. Figure 4 and Figure 6 The operations and / or methods of the point cloud encoder (or octree analyzer 40002) described above are the same or similar to the operations and / or methods. Figures 1 to 9 Same as those described.
[0208] According to an embodiment, the surface model processor 12004 may perform triplet geometry encoding based on the surface model to reconstruct the point position in a specific area (or node) based on voxels. Figure 4 The operations and / or methods described herein are the same or similar to the operations and / or methods of the point cloud encoder (e.g., surface approximation analyzer 40003). Figures 1 to 9 Same as those described.
[0209] According to the embodiment, the intra-frame / inter-frame encoding processor 12005 may perform intra-frame / inter-frame encoding on the point cloud data. The intra-frame / inter-frame encoding processor 12005 may perform the same as the reference. Figure 7 The intra / inter coding described is the same or similar to the coding. Details and references Figure 7 The intra / inter encoding processor 12005 may be included in the arithmetic encoder 12006 according to an embodiment.
[0210] According to an embodiment, the arithmetic encoder 12006 performs entropy coding on the octree and / or approximate octree of the point cloud data. For example, the coding scheme includes arithmetic coding. The arithmetic encoder 12006 performs the same or similar operations and / or methods as the operations and / or methods of the arithmetic encoder 40004.
[0211] The metadata processor 12007 according to an embodiment processes metadata (e.g., set values) about the point cloud data and provides it to necessary processing processes such as geometry coding and / or attribute coding. In addition, the metadata processor 12007 according to an embodiment may generate and / or process signaling information related to geometry coding and / or attribute coding. The signaling information according to an embodiment may be encoded separately from the geometry coding and / or attribute coding. The signaling information according to an embodiment may be interleaved.
[0212] The color conversion processor 12008, the attribute conversion processor 12009, the prediction / lifting / RAHT conversion processor 12010, and the arithmetic encoder 12011 perform attribute encoding. Figures 1 to 9The attribute codes described are the same or similar, so their detailed description is omitted.
[0213] The color transform processor 12008 according to an embodiment performs color transform encoding to transform the color value included in the attribute. The color transform processor 12008 may perform color transform encoding based on the reconstructed geometry. The reconstructed geometry is compared with the reference Figures 1 to 9 In addition, its execution is the same as that of reference Figure 4 The operations and / or methods of the color converter 40006 described above are the same as or similar to the operations and / or methods described above, and their detailed description is omitted.
[0214] The attribute transformation processor 12009 according to an embodiment performs attribute transformation to transform attributes based on reconstructed geometry and / or locations where geometry encoding is not performed. The attribute transformation processor 12009 performs the attribute transformation with reference to Figure 4 The operations and / or methods of the attribute converter 40007 described in the foregoing description are the same as or similar to those of the attribute converter 40007. Detailed description thereof is omitted. The prediction / lifting / RAHT transform processor 12010 according to the embodiment may encode the transformed attribute by any one or a combination of RAHT coding, prediction transform coding, and lifting transform coding. The prediction / lifting / RAHT transform processor 12010 performs the same as the reference. Figure 4 The operations of the RAHT transformer 40008, the LOD generator 40009 and the lifting transformer 40010 described above are the same or similar to at least one operation. In addition, the prediction transform coding, the lifting transform coding and the RAHT transform coding are the same as those of the reference Figures 1 to 9 Those described are the same, so their detailed description is omitted.
[0215] The arithmetic encoder 12011 according to the embodiment may encode the encoded attribute based on arithmetic coding. The arithmetic encoder 12011 performs the same or similar operation and / or method as that of the arithmetic encoder 400012.
[0216] The transmission processor 12012 according to the embodiment may send individual bitstreams containing encoded geometry and / or encoded attribute and metadata information, or send a bitstream configured with encoded geometry and / or encoded attribute and metadata information. When the encoded geometry and / or encoded attribute and metadata information according to the embodiment is configured as a bitstream, the bitstream may include one or more sub-bitstreams. The bitstream according to the embodiment may include signaling information and slice data, and the signaling information includes a sequence parameter set (SPS) for sequence level signaling, a geometry parameter set (GPS) for geometry information coding signaling, an attribute parameter set (APS) for attribute information coding signaling, and a patch parameter set (TPS) for patch level signaling. Slice data may include information about one or more slices. A slice according to the embodiment may include a geometry bitstream Geom00 and one or more attribute bitstreams Attr00 and Attr10.
[0217] A slice is a sequence of syntax elements that represents all or part of a coded point cloud frame.
[0218] The TPS according to an embodiment may include information about each of the one or more tiles (e.g., coordinate information and height / size information about a bounding box). The geometry bitstream may include a header and a payload. The header of the geometry bitstream according to an embodiment may include a parameter set identifier (geom_parameter_set_id), a tile identifier (geom_tile_id), and a slice identifier (geom_slice_id) included in the GPS, as well as information about the data contained in the payload. As described above, the metadata processor 12007 according to an embodiment may generate and / or process signaling information and send it to the transmission processor 12012. According to an embodiment, an element that performs geometry encoding and an element that performs attribute encoding may share data / information with each other as indicated by the dotted lines. The transmission processor 12012 according to an embodiment may perform operations and / or transmission methods that are the same or similar to those of the transmitter 10003. Details and References Figure 1 and Figure 2 Those described are the same, so their description is omitted.
[0219] Figure 13 A receiving device according to an embodiment is shown.
[0220] Figure 13 The receiving device shown is Figure 1 The receiving device 10004 (or Figure 10 and Figure 11 Example of a point cloud decoder. Figure 13 The receiving device shown can execute and refer to Figures 1 to 11One or more operations and methods that are the same or similar to those of the described point cloud decoder.
[0221] The receiving device according to the embodiment includes a receiver 13000, a receiving processor 13001, an arithmetic decoder 13002, an octtree reconstruction processor based on an occupancy code 13003, a surface model processor (triangle reconstruction, upsampling, voxelization) 13004, an inverse quantization processor 13005, a metadata parser 13006, an arithmetic decoder 13007, an inverse quantization processor 13008, a prediction / lifting / RAHT inverse transform processor 13009, a color inverse transform processor 13010 and / or a renderer 13011. Each decoding element according to the embodiment may perform an inverse process of the operation of the corresponding encoding element according to the embodiment.
[0222] The receiver 13000 according to the embodiment receives point cloud data. The receiver 13000 may perform the same Figure 1 The operation and / or receiving method of the receiver 10005 is the same as or similar to the operation and / or receiving method of the receiver 10005. The detailed description thereof is omitted.
[0223] The reception processor 13001 according to an embodiment may obtain a geometry bitstream and / or an attribute bitstream from the received data. The reception processor 13001 may be included in the receiver 13000.
[0224] The arithmetic decoder 13002, the octtree reconstruction processor 13003 based on the occupancy code, the surface model processor 13004, and the inverse quantization processor 13005 may perform geometric decoding. Figures 1 to 10 The geometric decoding described is the same or similar, so its detailed description is omitted.
[0225] The arithmetic decoder 13002 according to the embodiment may decode the geometry bitstream based on arithmetic coding. The arithmetic decoder 13002 performs the same or similar operations and / or encoding as those of the arithmetic decoder 11000.
[0226] The octree reconstruction processor 13003 based on the occupancy code according to the embodiment can reconstruct the octree by obtaining the occupancy code from the decoded geometry bitstream (or information about the geometry obtained as a result of decoding). The octree reconstruction processor 13003 based on the occupancy code performs the same or similar operations and / or methods as the operations of the octree synthesizer 11001 and / or the octree generation method. When triplet geometry coding is applied, the surface model processor 13004 according to the embodiment can perform triplet geometry decoding and related geometry reconstruction (e.g., triangle reconstruction, upsampling, voxelization) based on the surface model method. The surface model processor 13004 performs the same or similar operations as the operations of the surface approximation synthesizer 11002 and / or the geometry reconstructor 11003.
[0227] The inverse quantization processor 13005 according to an embodiment may inversely quantize the decoded geometry.
[0228] According to an embodiment, the metadata parser 13006 may parse metadata (eg, setting values) contained in the received point cloud data. The metadata parser 13006 may pass the metadata to the geometry decoding and / or attribute decoding. Figure 12 The metadata described is the same, so a detailed description thereof is omitted.
[0229] The arithmetic decoder 13007, the inverse quantization processor 13008, the prediction / lifting / RAHT inverse transform processor 13009 and the color inverse transform processor 13010 perform attribute decoding. Figures 1 to 10 The attribute decoding described is the same or similar, so the detailed description thereof is omitted.
[0230] The arithmetic decoder 13007 according to the embodiment may decode the attribute bitstream by arithmetic coding. The arithmetic decoder 13007 may decode the attribute bitstream based on the reconstructed geometry. The arithmetic decoder 13007 performs the same or similar operation and / or encoding as that of the arithmetic decoder 11005.
[0231] The inverse quantization processor 13008 according to an embodiment may inversely quantize the decoded attribute bitstream. The inverse quantization processor 13008 performs the same or similar operation and / or method as the inverse quantizer 11006 and / or the inverse quantization method.
[0232] The prediction / lifting / RAHT inverse transformer 13009 according to an embodiment may process the reconstructed geometry and inverse quantized attributes. The prediction / lifting / RAHT inverse transformer 13009 performs one or more operations and / or decoding that are the same or similar to the operations and / or decoding of the RAHT transformer 11007, the LOD generator 11008, and / or the inverse transformer 11009. The color inverse transformer 13010 according to an embodiment performs inverse transform encoding to inversely transform the color value (or texture) included in the decoded attribute. The color inverse transformer 13010 performs the same or similar operations and / or inverse transform encoding as the operations and / or inverse transform encoding of the color inverse transformer 11010. The renderer 13011 according to an embodiment may render point cloud data.
[0233] Figure 14 An exemplary structure that can be operated in conjunction with a point cloud data transmitting / receiving method / apparatus according to an embodiment is shown.
[0234] Figure 14 The structure of represents a configuration in which at least one of the server 1460, the robot 1410, the self-driving vehicle 1420, the XR device 1430, the smart phone 1440, the home appliance 1450, and / or the head mounted display (HMD) 1470 is connected to the cloud network 1400. The robot 1410, the self-driving vehicle 1420, the XR device 1430, the smart phone 1440, or the home appliance 1450 is referred to as a device. In addition, the XR device 1430 may correspond to a point cloud data (PCC) device according to an embodiment or may be operably connected to a PCC device.
[0235] The cloud network 1400 may represent a network that constitutes a part of a cloud computing infrastructure or exists in a cloud computing infrastructure. Here, the cloud network 1400 may be configured using a 3G network, a 4G or Long Term Evolution (LTE) network, or a 5G network.
[0236] The server 1460 may be connected to at least one of the robot 1410 , the self-driving vehicle 1420 , the XR device 1430 , the smart phone 1440 , the home appliance 1450 , and / or the HMD 1470 via the cloud network 1400 , and may assist in at least a portion of the processing of the connected devices 1410 to 1470 .
[0237] The HMD 1470 represents one of implementation types of an XR device and / or a PCC device according to an embodiment. The HMD type device according to an embodiment includes a communication unit, a control unit, a memory, an I / O unit, a sensor unit, and a power supply unit.
[0238] Hereinafter, various embodiments of devices 1410 to 1450 to which the above-described technology is applied will be described. Figure 14The illustrated devices 1410 to 1450 may be operably connected / coupled to the point cloud data transmitting and receiving devices according to the above-described embodiments.
[0239] <PCC+XR>
[0240] The XR / PCC device 1430 may employ PCC technology and / or XR (AR+VR) technology, and may be implemented as an HMD, a head-up display (HUD) disposed in a vehicle, a television, a mobile phone, a smart phone, a computer, a wearable device, a home appliance, a digital signage, a vehicle, a fixed robot, or a mobile robot.
[0241] The XR / PCC device 1430 may analyze 3D point cloud data or image data acquired through various sensors or from an external device and generate position data and attribute data about 3D points. Thus, the XR / PCC device 1430 may acquire information about the surrounding space or real objects, and render and output an XR object. For example, the XR / PCC device 1430 may match an XR object including auxiliary information about the recognized object with the recognized object and output the matched XR object.
[0242] <PCC+Autopilot+XR>
[0243] The self-driving vehicle 1420 may be realized as a mobile robot, a vehicle, an unmanned aerial vehicle, etc. by applying PCC technology and XR technology.
[0244] The self-driving vehicle 1420 to which the XR / PCC technology is applied may represent a self-driving vehicle provided with a means for providing an XR image, or a self-driving vehicle as a control / interaction target in the XR image. Specifically, as a control / interaction target in the XR image, the self-driving vehicle 1420 may be distinguished from the XR device 1430 and may be operably connected thereto.
[0245] The self-driving vehicle 1420 having means for providing XR / PCC images may acquire sensor information from sensors including cameras, and output generated XR / PCC images based on the acquired sensor information. For example, the self-driving vehicle 1420 may have a HUD and output XR / PCC images thereto, thereby providing passengers with XR / PCC objects corresponding to real objects or objects presented on the screen.
[0246] When the XR / PCC object is output to the HUD, at least a portion of the XR / PCC object may be output to overlap with the real object to which the passenger's eyes are directed. On the other hand, when the XR / PCC object is output on a display provided in the self-driving vehicle, at least a portion of the XR / PCC object may be output to overlap with the object on the screen. For example, the self-driving vehicle 1220 may output XR / PCC objects corresponding to objects such as a road, another vehicle, a traffic light, a traffic sign, a two-wheeled vehicle, a pedestrian, and a building.
[0247] Virtual reality (VR) technology, augmented reality (AR) technology, mixed reality (MR) technology, and / or point cloud compression (PCC) technology according to an embodiment are applicable to various devices.
[0248] In other words, VR technology is a display technology that only provides CG images of real-world objects, backgrounds, etc. On the other hand, AR technology refers to a technology that displays a virtually created CG image on an image of a real object. MR technology is similar to the above-mentioned AR technology in that the virtual objects to be displayed are mixed and combined with the real world. However, MR technology is different from AR technology in that AR technology clearly distinguishes between real objects and virtual objects created as CG images and uses virtual objects as supplementary objects to real objects, while MR technology treats virtual objects as objects with properties equivalent to real objects. More specifically, an example of the application of MR technology is a holographic service.
[0249] Recently, VR, AR, and MR technologies are often referred to as extended display (XR) technologies rather than being clearly distinguished from each other. Therefore, embodiments of the present disclosure are applicable to any of VR, AR, MR, and XR technologies. Encoding / decoding based on PCC, V-PCC, and G-PCC technologies are applicable to such technologies.
[0250] The PCC method / apparatus according to the embodiment may be applied to a vehicle that provides a self-driving service.
[0251] Vehicles providing self-driving services are connected to the PCC device for wired / wireless communication.
[0252] When a point cloud data (PCC) transmitting / receiving device according to an embodiment is connected to a vehicle for wired / wireless communication, the device may receive / process content data related to an AR / VR / PCC service (which may be provided together with a self-driving service) and send it to the vehicle. In the case where the PCC transmitting / receiving device is installed on the vehicle, the PCC transmitting / receiving device may receive / process content data related to the AR / VR / PCC service and provide it to the user based on a user input signal input through a user interface device. The vehicle or user interface device according to the embodiment may receive a user input signal. The user input signal according to the embodiment may include a signal indicating a self-driving service.
[0253] Figure 15 The structure of geometry data and attribute data for spatial scalability according to an embodiment is shown.
[0254] Figure 15 The G-PCC decoder 15000 and / or the G-PCC decoder 15010 in may correspond to Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Decoding 20003 and Figure 10 and Figure 11 The decoder in Figure 13 The decoder and Figure 14 The XR device 1430 in. Figure 15 The G-PCC decoder 15000 and / or the G-PCC decoder 15010 in the embodiment may be executed Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 2 Code 20001, Figure 4 The encoder in Figure 12 The encoder and Figure 14 The reverse process of the operation on the sending side of the XR device 1430 in FIG.
[0255] According to an embodiment, the PCC data encoder may correspond to an encoding device, an encoder, a point cloud transmitting device, a point cloud data encoder, etc.
[0256] According to an embodiment, the geometric information may be referred to as geometric data, etc., and the attribute information may be referred to as attribute data, etc.
[0257] The method / apparatus according to an embodiment refers to a method / apparatus for transmitting / receiving point cloud data according to an embodiment.
[0258] Embodiments relate to an attribute compression method that supports spatial scalability of a geometry-based point cloud compression (G-PCC) encoder / decoder for 3D point cloud data compression.
[0259] The embodiment proposes a method for changing the point sampling technique of each LOD based on an octree during attribute encoding / decoding of G-PCC to provide spatial scalability.
[0260] Specifically, 1) a method for reorganizing points through octree nodes using Morton codes, 2) a method for processing isolated points, 3) a point sampling technique for each node, and 4) a signaling method that supports the above three methods are proposed.
[0261] When point cloud data is dense, it may be desirable to use narrower bandwidth (e.g., thumbnails) or to send and process lower resolution point clouds to reduce computational complexity at the decoder side. This capability is called spatial scalability.
[0262] In geometry encoding / decoding, a spatial scalability function may be provided by a process of encoding or decoding occupied bits only up to a depth level selected by adjusting the depth level of an octree.
[0263] The attribute encoding / decoding according to the embodiment may perform an operation based on an octree structure according to the embodiment to provide a spatial scalability function. A method of generating LOD from a selected depth level of an octree, a method of calculating neighbor nodes, and a method of calculating weights according to the embodiment may be required to provide spatial scalability in attribute encoding / decoding.
[0264] The attribute encoding / decoding according to the embodiment may be based on the technique of changing the sampling points of each LOD based on the octree and configuring the most suitable points of each LOD as nodes to provide spatial scalability. Thus, the attribute compression efficiency may be enhanced.
[0265] The embodiment proposes a method for changing the point sampling technique of each LOD based on an octree to provide spatial scalability in the attribute encoding / decoding process of G-PCC.
[0266] The LOD sampling-based method according to an embodiment may refer to a method of sorting points with a Morton code and classifying the points selected according to the embodiment into lower LODs that can be candidates for neighbor sets. The LOD sampling-based method according to an embodiment may increase the accuracy and compression efficiency of neighbor point set generation while reducing the complexity of LOD generation.
[0267] like Figure 15As shown, PCC data (or point cloud data) may be composed of a geometry bit stream and an attribute bit stream. The point cloud data may be encoded and transmitted by a point cloud data transmitting device or a point cloud data encoder according to an embodiment. A point cloud data receiving device or a point cloud data decoder according to an embodiment may decode the point cloud data.
[0268] In this case, for geometry data (or geometry bitstream) and / or attribute data (or attribute bitstream), the cloud data receiving device may partially decode an octree-structured geometry bitstream and / or an attribute bitstream based on lifting coding in the order of detail levels (LOD).
[0269] The G-PCC decoder 15000 and / or the G-PCC decoder 15010 represent a decoder of the point cloud data receiving device according to an embodiment. Specifically, the G-PCC decoder 15000 and / or the G-PCC decoder 15010 represent a geometry decoder and / or an attribute decoder of a PCC decoder.
[0270] The G-PCC decoder 15000 may not decode all of the geometry bitstream and / or attribute bitstream in the received bitstream, but may partially decode the geometry bitstream and / or attribute bitstream to provide low-resolution point cloud data to the user.
[0271] The G-PCC decoder 15010 may decode all geometry bitstreams and / or attribute bitstreams in the received bitstream and provide the decoded bitstream to a user.
[0272] The PCC encoder on the transmitting side corresponding to the G-PCC decoder 15000 and the G-PCC decoder 15010 may encode and transmit the entire point cloud data, or may encode and transmit part of the point cloud data.
[0273] Figure 16 The attribute encoding / decoding process according to an embodiment is shown.
[0274] Figure 16 The process may be applied to encoding and / or decoding according to an embodiment. For example, the process may be Figure 15 Decoders 15000 and 15010 and / or corresponding encoders ( Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 2 Code 20001, Figure 4 The encoder in Figure 12 The encoder and Figure 14 Executed by the XR device 1430 in .
[0275] Changes and combinations between the embodiments of the present disclosure are possible.In addition, the terms used in the present disclosure may be understood based on the intended meanings of the terms within the scope widely adopted in the relevant art.
[0276] The attribute encoder / decoder according to the embodiment may apply the octree structure as in the geometry encoding / decoding to provide a spatial scalability function when encoding / decoding the attribute. The embodiment includes a method of reorganizing points belonging to a node starting from a selected depth level of the octree and a method of sampling points of each LOD.
[0277] The method of reorganizing points and the method of sampling points of each LOD are performed in PCC attribute encoding / decoding of a PCC decoder / encoder, and can be applied when a predictive transform technique or a lifting transform technique is used in the attribute encoding / decoding process.
[0278] S16000 sorts point cloud data based on Morton order
[0279] Points in the point cloud may be converted to Morton codes based on x, y, and z position values and may be sorted based on the Morton code values.
[0280] S16010 Set partial encoding / decoding area of attribute data
[0281] In attribute encoding / decoding, as in geometry encoding / decoding (see Figure 17 ), the depth level of the octree can be set, and the points to which attribute encoding (decoding) is to be applied can be constructed and encoded / decoded starting from the set octree depth level.
[0282] S16020 uses Morton codes to reorganize points belonging to octree nodes
[0283] According to the embodiment, the method for reorganizing the points sorted by the Morton code for each octree node may adopt the pseudo code (formula) given below.
[0284] nodeSizeLog2: indicates the depth level of the octree ( Figure 17 of 17000).
[0285] PX_mortoncode: represents the Morton code of node PX.
[0286] end_range_mc: indicates the range of the node's Morton code.
[0287] maxNumOfPoints: Indicates the maximum number of points for a particular octree depth level. This is the maximum number of points a node can ideally have.
[0288] Int maxNumOfPoints=pow(8,nodeSizeLog2+1);
[0289] For(int I=1;I <indexCount;++i){
[0290] if(Pi_mortoncode <end_range_mc){
[0291] …
[0292] }else{
[0293] end_range_mc=Pi_mortoncode+
[0294] (maxNumOfPoints-(Pi_mortoncode%maxNumOfPoints));
[0295] }
[0296] }
[0297] According to the above syntax, the range (end_range_mc) with the same Morton code value within a node in the octree structure can be known, and the node can be selected and uploaded to a higher node.
[0298] The same meaning as the above syntax can be expressed as follows.
[0299] When using the boost method, the following node selection methods are available.
[0300] When lifting_scalability_enabled_flag is equal to 1 (lifting is used), the following applies:
[0301] endIndex=0
[0302] assignedPointCount=0
[0303] for(lod=minGeomNodeSizeLog2; unprocessedPointCount>0; lod++){
[0304] startIndex=assignedPointCount
[0305] nonAssignedPointCount=0
[0306] samplingFromLastFlag=lod&1
[0307] for(i=0;i<unprocessedPointCount;i++){
[0308] currVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i]]>>(3*×(lod+1))
[0309] if(samplingFromLastFlag==0){
[0310] if(i==0)
[0311] nonAssignedPointIndexes[nonAssignedPointCount++]=unprocessedPointIndexes[i]
[0312] else{
[0313] prevVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i-1]]>>(3*×(lod+1))
[0314] if(currVoxelIndex>prevVoxelIndex)
[0315] nonAssignedPointIndexes[nonAssignedPointCount++]=unprocessedPointIndexes[i]
[0316] else
[0317] assignedPointIndexes[assignedPointCount++]=unprocessedPointIndexes[i]
[0318] }
[0319] }else{
[0320] if(i==(unprocessedPointCount–1))
[0321] nonAssignedPointIndexes[nonAssignedPointCount++]=unprocessedPointIndexes[i]
[0322] else{
[0323] nextVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i+1]]>>(3*×(lod+1))
[0324] if(currVoxelIndex<nextVoxelIndex)
[0325] nonAssignedPointIndexes[nonAssignedPointCount++]=unprocessedPointIndexes[i]
[0326] else
[0327] assignedPointIndexes[assignedPointCount++]=unprocessedPointIndexes[i]
[0328] }
[0329] }
[0330] }
[0331] endIndex=assignedPointCount
[0332] if(startIndex!=endIndex){
[0333] numOfPointInSkipped=PointNumInSlice–PointCount
[0334] if((endIndex–startIndex)>(startIndex+numOfPointInSkipped)){
[0335] for((loop=0;loop<lod-minGeomNodeSizeLog2;loop++)){
[0336] computeNearestNeighbours(
[0337] PointCount-unprocessedPointCountPerLevelOfDetail[loop],
[0338] PointCount-unprocessedPointCountPerLevelOfDetail[loop+1],
[0339] loop+minGeomNodeSizeLog2,assignedPointIndexes,
[0340] McodeUnsorted,nonAssignedPointCount,
[0341] nonAssignedPointIndexes))
[0342] }
[0343] }
[0344] }
[0345] computeNearestNeighbours(
[0346] startIndex,endIndex,
[0347] lod,assignedPointIndexes,
[0348] McodeUnsorted,nonAssignedPointCount,
[0349] nonAssignedPointIndexes)
[0350] unprocessedPointCountPerLevelOfDetail[lod+1]=nonAssignedPointCount
[0351] unprocessedPointCount=nonAssignedPointCount
[0352] unprocessedPointIndexes=nonAssignedPointIndexes
[0353] }
[0354] Specifically, when the next voxel index is set to nextVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i+1]]>>(3*×(lod+1)), the Morton code may be shifted to determine whether the Morton code values are within the same range.
[0355] The encoder according to an embodiment may estimate end_range_mc from the Morton code of the point. When the Morton code value of the next point is within the range end_range_mc, the points belong to the same node. When the Morton code has a larger value outside the range, the encoder may estimate end_range_mc based on the Morton code of the point when the node is changed to a new node, and apply it to the subsequent points. In this way, points belonging to the octree node can be reorganized between the points sorted by the Morton code (see Figure 18 ).
[0356] S16030 Merge isolated nodes
[0357] Nodes can be merged when the number of points belonging to a node is less than a certain threshold (see Figure 18 ).
[0358] For example, when the minimum number of points of a node is "a" and the number of points is less than the minimum number of points, the node may be merged with the next node. In addition, the value of a may be signaled. When the maximum merge range is r, at most the next r nodes may be merged, and the value of r may be signaled. When the number of points is less than the minimum number of points, the node may be merged with the next node. When the number of merged points is less than the minimum number of points, the node may be merged with the next node. This repetition may be performed r times.
[0359] S16040 configures LOD of octree nodes
[0360] Depending on the implementation, LOD configuration may proceed from the root node to the leaf node, or may proceed from the leaf node to the root node.
[0361] For example, when configuring LOD from root to leaf, LOD configuration may be performed from root node to leaf node. Consistency may be maintained. When configuring LOD from leaf to root, LOD configuration may be performed from leaf node to root node.
[0362] S16050 performs LOD sampling of octree nodes
[0363] Sampling may be performed by selecting points from among the points belonging to a node (including a merge node according to an embodiment). Attribute compression may be affected by the selected points. The best reference points are selected and reorganized into a reference region (retained list), and a set of neighbor nodes is constructed from the reference region. Therefore, compression efficiency may be affected.
[0364] The LOD configuration unit 20120 according to an embodiment may use the following three sampling methods.
[0365] <Select point N>
[0366] When selecting a point among points belonging to a node (including a merge node), the Nth point may be selected. When compression is performed by an encoder, N may be input by a user, or an optimal value may be found in an RDO (Rate Distortion Optimization) manner.
[0367] The value of N may be sent to a decoder (receiving device) as a parameter (metadata or signaling data). When the value of N is not specified, a value of 0 (ie, a first value) may be selected.
[0368] The compression efficiency may depend on which point is chosen.
[0369] <Select a point close to the center of the octree parent node area>
[0370] When selecting a point among the points belonging to a node (including a merged node), the point closest to the center of the bounding box of the parent node may be selected. When the selected point is closer to the center position of the parent node, the error may be further reduced when selecting neighbor points and calculating the weights of the neighbor points.
[0371] When selecting neighbor points, the distance between points is calculated, and points with closer distance can be selected as neighbor points. In order to provide spatial scalability function, the position value of a point can be replaced by the value of the center position of the node to which the point belongs (the node can change according to the level of LOD). In this state, the distance can be calculated. Therefore, in order to find the neighbor point of point Pi, the position value of Pi can be changed to the center position of the parent node according to LOD1, and the position of the point Ri in the neighbor point candidate list (reserved list) R is also changed according to LOD1. Then, the distance of the point can be calculated. In other words, the distance between the center positions of nodes of the same depth level can be calculated, so that the neighbor node is set as the neighbor point. From the distance calculated in this way, the weight of the neighbor point can be obtained, 1 / distance=Pi. Since the distance is calculated by changing the position of the point to the center position of the parent node, the calculation contains errors. As the depth level of the node decreases, that is, the bounding box of the node is larger, the error can increase.
[0372] In order to reduce the error, the transmitting / receiving device and method according to the embodiment can select a point close to the center position of the parent node area when selecting a reference point in encoding or decoding. The error can be reduced according to the position value of the reference point, and the reduced error can have a good effect on the compression efficiency.
[0373] <Select the point with the middle index>
[0374] When selecting a point from among the points belonging to a node (including a merged node), a point with an intermediate index from among the points may be selected. Since the node with the intermediate index has a high probability of being close to the center position of the bounding box of the parent node (in many cases), and selecting a point by the intermediate index does not require much additional calculation, the probability of finding a point close to the center position with a small amount of calculation may be high.
[0375] Figure 17 An octree structure of point cloud data according to an embodiment is shown.
[0376] Figure 17 Show Figure 16 Operation S16010 in .
[0377] When the point cloud transmitting / receiving device or the encoder / decoder according to the embodiment encodes or decodes the attribute, it can be as follows Figure 17 The octree structure is used as shown in . The octree structure is interpreted in the order from the root node to the leaf node, that is, from the upper node (parent node) to the lower node (child node), and the nodes can be arranged from left to right in Morton order. In addition, the LOD can be configured (generated) in the order from the leaf node to the root node.
[0378] For example, an octree depth level such as a starting depth of 17000 is set, and the G-PCC decoder 15000 may be as follows: Figure 15 As shown, the point cloud data of the set octree depth level 17000 is decoded.
[0379] A parent node may have 8 or fewer nodes as child nodes of the parent node. In the process of sampling a specific node among the child nodes and shifting it to the parent node, the frontmost point 17050 among the points 17040 sorted by the Morton code belonging to the parent node may be selected as a reference node or a retained node (see Figure 17 ).
[0380] Specifically, in order to support spatial scalability, the transmitting device or encoder according to the embodiment may encode and transmit all points, or encode and transmit points of a specific depth. Similarly, the receiving device or decoder according to the embodiment may decode all points or points of a specific depth.
[0381] like Figure 17 As shown, when there is an octree representing points of the received point cloud data, a process of spatial scalability decoding based on a starting depth of 17000 will be described.
[0382] When the level of the octree is 0 (17010), level 1 (17020), and level 2 (17030), a starting depth 1730 may be set, points may be decoded starting from the starting depth 1730, and points below the starting depth may be ignored.
[0383] Starting from the point at the starting depth 1730, the points may be reorganized and merged to configure the LOD, and sampled. Then, the transmitting device may encode the attribute, or the receiving device may decode the attribute.
[0384] Figure 18A method of reorganizing points belonging to a Morton code based node according to an embodiment is shown.
[0385] Figure 18 Show Figure 16 Operation S16020 in .
[0386] According to an embodiment, the encoder may be based on Figure 16 Operation S16020 in reorganizes the points of the child node 18010 belonging to the parent node 18000.
[0387] Based on Figure 17 After setting the attribute encoding / decoding area at the starting depth 17000 in, the points of the octree belonging to the level at or after the starting depth 17000 can be as follows Figure 18 The order of the levels in Figure 16 The formula in S16020 is reorganized.
[0388] For example, Figure 17 The points of the level (depth) of the octree set in may be P0, P1, P2, ..., P13, .... In order to ensure efficient encoding / decoding, the method / apparatus (including encoder / decoder) according to the embodiment may reorganize the points into groups belonging to the same category, such as Figure 18 shown.
[0389] according to Figure 16 The formula "end_range_mc = Pi_mortoncode + (maxNumOfPoints - (Pi_mortoncode% maxNumOfPoints))" is used. The value of end range mc can be calculated using the octree level value of P0 and the Morton code value (e.g., 1). When the first range value is 8, the Morton code value MC of each of P0 to P5 is less than or equal to 8, so the points can be reorganized into points belonging to the same node (18010).
[0390] The second range value may use the octree level value of P6, the Morton code value (e.g., 9) and Figure 16 When the point whose Morton code value is less than the second range value is only P6, P6 is reorganized into a single point (18020).
[0391] The third range value may use the octree level value or Morton code value of P7 (e.g., 16) and Figure 16 P7 to P10 having Morton code values included in the third range of values may be reorganized as points belonging to the same node (18030).
[0392] The fourth range value may use the octree level value or Morton code value of P11 (e.g., 26) and Figure 16Points having Morton code values included in the fourth range of values may be reorganized into points belonging to the same node (18040).
[0393] Figure 19 Shown is merging isolated nodes according to an embodiment.
[0394] Figure 19 Show Figure 16 Example of S16030 in .
[0395] Isolated nodes are those whose number upon reorganization is less than or equal to a certain number.
[0396] As an example, the following will be described Figure 16 The minimum number of node points shown in S16030 is 4.
[0397] For example, there may be points 19000 sorted in the order of the Morton code. Figures 16 to 18 As a result of the described operation of reorganizing the ordered points, the number of points belonging to the same node, such as in 19010 , may be greater than 4. In addition, there may be points with 4 or fewer points belonging to the same node, such as in 19020 and 19030 .
[0398] The transmission / reception method / apparatus according to the embodiment may merge points whose number is less than or equal to a certain number, rather than leaving the points as separate nodes. According to the embodiment, the merging process may be optional.
[0399] As isolated nodes are merged, the number of recombinant points may be greater than or equal to a specific number. Recombinant points belonging to the same node have a sampling range. In addition, the merged node has a merged sampling range.
[0400] Figure 19 An example in which a node having one point and a node having four points are merged and an example 19040 in which a node having three points and a node having two points are merged are shown.
[0401] Figure 20 A PCC data encoder according to an embodiment is shown.
[0402] Figure 20 is an encoder according to an embodiment ( Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 2 Code 20001, Figure 4 The encoder in Figure 12 The encoder and Figure 14 1430), which executes Figure 16 Operations in .
[0403] The PCC data encoder according to the embodiment may be referred to as an encoder, and the PCC data decoder according to the embodiment may be referred to as a decoder.
[0404] The PCC data (point cloud data) may be input to and encoded by the encoder, and may output a geometry information bit stream and an attribute information bit stream (a bit stream for point cloud data).
[0405] Figure 20 The various elements in the description may correspond to hardware, software, a processor and / or a combination thereof.
[0406] The data input unit 20000 may receive geometric data, attribute data and / or parameters related thereto.
[0407] The coordinate converter 20010 may convert coordinates related to the position (coordinate) information of the geometric data. The coordinate converter 17010 may correspond to Figure 4 The coordinate transformer 40000 in.
[0408] The geometric information transform quantization processor 20020 may be referred to as a geometric information transform quantizer. The geometric information transform quantization processor 20020 may correspond to Figure 4 Quantizer 40001 in. The geometric information transform quantization processor 20020 can receive a full-scale value (quantization value) and can adjust the scale of each axis according to the distribution of the content.
[0409] The spatial segmenter 20030 may segment the point cloud data into spatial segmentation units according to the implementation. The spatial segmenter 20030 may perform spatial adaptive segmentation of the data space using tiles, blocks, density-based methods, axis-based methods, attribute-based methods, geometry-based methods, etc. according to the data characteristics of the point cloud data. The spatial segmenter 20030 may perform data segmentation. A frame may be segmented into tiles and slices. The source point cloud data may be segmented into multiple slices and encoded in a bitstream. A slice is a collection of points that can be encoded or decoded independently. A slice may include a geometry data unit and zero or one or more attribute data units. The attribute data unit may depend on the corresponding geometry data in the same slice. In a slice, a geometry data unit may appear before a related attribute unit. The data units in a slice are adjacent. A group of slices may be identified by a common tile identifier. Tile information describing the bounding box of each tile may be present in the bitstream. A tile may overlap with other tiles in the bounding box. Each slice may include an index to which the tile belongs.
[0410] The geometry information encoder 20040 can encode the geometry data of the point cloud data. The detailed process of geometry encoding is described below.
[0411] The voxelization processor 20050 may voxelize the geometric data. The voxelization processor 20050 may correspond to Figure 4 Voxelizer 40001 in .
[0412] The octree generator 20060 may be referred to as an octree generation unit. The octree generator 20060 may correspond to Figure 4 The octree analyzer 40002 in. The octree generator 20060 can generate geometric data according to the octree structure.
[0413] When using a geometric octree, geometric encoding can be performed as follows: 1) A bounding box defined by two points can be defined. The octree structure is constructed by recursively subdividing the bounding box. Each cube is divided into 8 sub-cubes. The 8-bit code can be called an occupancy code. One bit per sub-cube can be represented as 1 when the point is filled, or as 0 when the point is empty. Sub-cubes with sides greater than 1 are further divided. Sub-cubes of size 1 can be voxels. Multiple points can overlap and map to the same sub-cube (voxel). The points of each sub-cube can be arithmetically encoded.
[0414] The decoding process can be started by reading the size of the bounding box from the bitstream. The same octree structure can be generated by dividing the bounding box according to the occupancy code. When a sub-cube of size 1 is reached, the points of the sub-cube can be arithmetically decoded.
[0415] The octree generator 20060 constructs an octree from the received geometry information bitstream (geometry data). For the spatial scalability function, the octree generator 20060 may construct geometry information only at a specific level of the octree. In this case, the position value may be changed to the value of the center of the bounding box.
[0416] The octree generator 20060 may construct and decode points to which attribute decoding is to be applied for partial geometry decoding starting from the restored octree depth level.
[0417] The geometry information predictor 20070 may perform prediction on the geometry data. The geometry information predictor 20070 may generate a prediction value for the geometry data between the parent node and the child node in the octree structure. A residual bit value may be generated by the prediction value to perform entropy coding on the residual value to construct a geometry bitstream. Efficient bitstream transmission may be achieved by encoding only the residual value for the prediction value. The geometry information predictor 20070 may recover (reconstruct) the geometry data from the geometry octree for attribute coding. This is because the attribute coding is performed based on the reconstructed geometry data.
[0418] The geometry information entropy encoder 20080 may perform entropy encoding on the geometry data to generate a geometry bitstream.
[0419] The attribute information encoder 20090 may encode the attribute data as follows.
[0420] The color conversion processor 20100 may convert the color of the attribute data according to the characteristics of the data.
[0421] The color rescaler 20110 may be referred to as a recoloring unit, a color recolorer, or a recolorer. According to an embodiment, when overlapping points are merged in the reconstructed geometry octree structure, the attribute information encoder 20090 may perform recoloring, or perform RAHT attribute encoding or configure LOD according to prediction / lifting transform attribute encoding. According to an embodiment, when overlapping points are not merged, recoloring and / or LOD configuration may be performed even in RATH encoding. According to an embodiment, when attribute information encoding is performed based on reconstructed geometry data, recoloring may be performed in the case of lossy geometry, or LOD configuration may be performed in the case of prediction / lifting transform attribute encoding instead of lossy geometry. When multiple points exist in one voxel and the overlapping points are merged to form the voxel by lossy compression, the operation of the color rescaler may be applied. In addition, when the geometry quantization coefficient is less than 1, multiple points may be included in one voxel. In this case, recoloring, LOD configuration, and neighbor point set configuration may be performed. The color rescaler 20110 may search for neighbor points close to the center position of the voxel in the original point cloud to set the representative color of the voxel, and calculate the representative color based on the neighbor points. In the process of searching for the neighbor points, the geometric information may be inverse quantized to restore the position difference between the original value and the quantized value. The color rescaler 20110 may perform inverse quantization for each axis when performing inverse quantization.
[0422] The LOD configuration unit 20120 may configure LOD from point cloud data. The LOD configuration unit 20120 may generate LOD from attribute data of re-adjusted color based on restored geometry information (geometry data). LOD may be generated when a prediction / lifting transform is applied during attribute encoding instead of RAHT. The LOD configuration unit 20120 may use Morton codes to reorganize points belonging to a node, merge isolated nodes, perform LOD sampling in an octree node according to an embodiment, and generate and transmit relevant information as signaling information.
[0423] The LOD configuration unit 20120 may set the depth level of the octree to support spatial scalability in attribute encoding / decoding, and configure and encode / decode points to which attribute encoding is to be applied starting from the set octree depth level (see Figure 17 ). The LOD configuration unit 20120 may be referred to as an LOD generator, an LOD configuration device, etc.
[0424] The octree depth level may be sent as signaling information by the encoder according to an embodiment to the decoder, and may be changed according to its purpose or the computational performance of the decoder.
[0425] The LOD configuration unit 20120 may perform point reorganization for each octree node by calculating the Morton code range of each octree node at the points sorted by the Morton code (see Figure 18 ).
[0426] The LOD configuration unit 20120 may set the minimum number of node points (the value of a) to distinguish the case where the number of points of the node is less than the minimum number of node points. The set value may be signaled to the decoder as signaling information.
[0427] The process of merging a node with the next consecutive node when the number of points of the node is less than the minimum number of points of the node can be called isolated node merging processing (see Figure 19 ), whether to perform isolated node merging processing can be notified to the decoder by signaling information.
[0428] When the number of points of a node is less than the minimum number of points of a node, the node can be merged with the next consecutive node. It is allowed to perform the merge up to r times (see Figure 19 ). The maximum number of merging times r can be signaled to the decoder as signaling information.
[0429] There are methods for selecting (sampling) a point from the points belonging to a node (including merging nodes) (see Figure 16 ). The method may be signaled to the decoder as signaling information. As the method, the Nth point may be selected, a point close to the center position of the octree parent node area may be selected, or a point with an intermediate index may be selected.
[0430] The neighbor point set configuration unit 20130 configures a neighbor point set for points belonging to the LOD in the LOD and higher LODs based on the LOD.
[0431] After the LOD1 set is generated, X (>0) nearest neighbors can be found in the group of the same or lower LOD (the group with longer distances between nodes) based on the LOD1 set, and registered in the predictor as a neighbor point set.
[0432] Each point can have a single predictor. Attributes are predicted from neighboring points registered with the predictor.
[0433] The predictor can be used to predict the attribute. The average value of the value obtained by multiplying the attribute of the registered neighbor point by the weight can be used as the prediction result, or a specific point can be used as the prediction result. As to which method to use, the compression result value can be calculated in advance, and then the method that can generate the minimum flow can be selected.
[0434] The residual between the attribute value of a point and the attribute value predicted by the predictor of the point may be encoded and signaled to the receiver along with the method selected by the predictor.
[0435] The decoder may decode the transmitted prediction method and predict the property value according to the method. The property value may be restored by decoding the transmitted residual value and adding the predicted value to it.
[0436] The attribute information predictor 20140 predicts the attribute value of a point from a set of neighboring points.
[0437] The residual attribute information quantization processor 20150 may generate a predicted value of the attribute data, generate a residual value between the predicted value and the attribute value, and quantize the residual value.
[0438] For example, when encoding a point Px, when configuring a set of neighbor points of Px, the attribute value of Px can be predicted based on the attribute values of the neighbor points in the set, the residual between the predicted attribute value and the attribute value of Px can be estimated, and the value of the residual can be encoded and sent to the decoder. The decoder configures a set of neighbor points of Px through LOD generation and the received additional information, predicts the attribute value of Px based on the attribute values of the neighbor points in the neighbor set, and restores the attribute value of Px by adding the received residual value to the predicted value.
[0439] When the prediction mode and residual information separation flag is true, the residual attribute information quantizer 20150 may calculate the residual between the predicted value according to the prediction mode and the value of the point, and encode its quantized value together with the prediction mode.
[0440] When the prediction mode and residual information separation flag is false, the residual attribute information quantizer 20150 may calculate the residual between the prediction value and the point value according to the prediction mode, and encode only the quantized value thereof.
[0441] The attribute information entropy encoder 20160 may perform entropy encoding on the attribute data and generate an attribute bitstream. For example, the residual value may be encoded and the attribute bitstream may be generated.
[0442] Figure 21 A PCC data decoder according to an embodiment is shown.
[0443] Figure 21 The decoder in can correspond to Figure 15 G-PCC decoder 15000, Figure 15 The G-PCC decoder 15010 in Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Decoding 20003,Figure 10 and Figure 11 The decoder in Figure 13 The decoder and Figure 14 The XR device 1430 in.
[0444] The geometry information bitstream and the attribute information bitstream encoded as input to the decoder may be input and decoded, and reconstructed PCC data may be output.
[0445] Figure 21 The various elements in the description may correspond to hardware, software, a processor and / or a combination thereof.
[0446] The geometry information decoder 210000 may receive the geometry information bitstream and decode the geometry information. The geometry information decoder 210000 may be a decoder on the receiving side corresponding to the geometry information encoder 20040 on the transmitting side. The geometry information decoder 210000 may perform the reverse process of the operation of the geometry information encoder 20040.
[0447] The geometry information entropy decoder 210010 may receive the geometry information bitstream and decode the geometry information based on an entropy scheme. The geometry information entropy decoder 210010 may perform the inverse process of the operation of the geometry information entropy encoder 20080.
[0448] The octree reconstructor 210020 may reconstruct the octree from the geometric information. The octree reconstructor 210020 may perform the reverse process of the operation of the octree generator 20060.
[0449] The geometry information predictor 210030 may generate a prediction value of the geometry information. For example, an inter-node prediction value may be generated in an octree structure, and the geometry information may be efficiently decoded using a residual value of the prediction value. The geometry information predictor 210030 may perform the inverse process of the operation of the geometry information predictor 20070.
[0450] The geometry information transform inverse quantization processor 210040 may inversely quantize the geometry data. For example, the geometry information transform inverse quantization processor 210040 may obtain the scale value (quantization value) of each axis from the signaling information (parameter) included in the bitstream, and may apply inverse quantization to the geometry information reconstructed based on the scale value of each axis.
[0451] The coordinate inverse transformer 210050 may inversely transform the coordinate information about the geometric information. The coordinate inverse transformer 210050 may perform an inverse process of the operation of the coordinate transformer 20010.
[0452] The attribute information decoder 210060 may receive the attribute information bitstream and decode the attribute information. The attribute information decoder 210060 may perform the reverse process of the operation of the attribute information encoder 20090.
[0453] The attribute residual information entropy decoder 210070 may decode the attribute residual information (eg, the residual value of the attribute information) contained in the attribute information bitstream based on an entropy scheme. The attribute residual information entropy decoder 210070 may perform the inverse process of the operation of the attribute information entropy encoder 20160.
[0454] The LOD configuration unit 210080 may configure the LOD from the point cloud data. The LOD configuration unit 210080 may generate the LOD from the attribute data of the re-adjusted color based on the restored geometry information (geometry data). When a prediction / lifting transform is applied during attribute encoding instead of RAHT, the LOD may be generated. The LOD configuration unit 210080 may correspond to Figure 20 LOD configuration unit 20120.
[0455] The LOD configuration unit 210080 may use Morton code to reorganize points belonging to a node, merge isolated nodes, and perform LOD sampling according to an embodiment in an octree node.
[0456] LOD configuration unit 210080 can construct and decode points to which attribute decoding is to be applied starting from the octree depth level recovered for partial attribute decoding (see Figure 17 ).
[0457] The LOD configuration unit 210080 can perform point reorganization for each octree node by calculating the Morton code range of each octree node from the points sorted by the Morton code (see Figure 18 ).
[0458] When the number of points of a node is less than the minimum number of restored node points (ie, a), the LOD configuration unit 210080 may merge the continuous nodes with the node (see Figure 19 ).
[0459] The process of merging a node with the next consecutive node when the number of points of the node is less than the minimum number of points of the node may be referred to as an isolated node merging process, and the merging may be performed according to the decoded indication of whether the isolated node merging process is to be performed.
[0460] When the number of points of a node is less than the minimum number of points of a node, the node may be merged with the next consecutive node. It may be allowed to perform the merge r times (ie, the maximum number of merges to be restored).
[0461] A point may be selected from among the points belonging to a node (including a merged node) based on a scheme for selecting a point to be restored. The scheme may include a scheme for selecting the Nth point, a scheme for selecting a point close to the center position of the octree parent node area, and a scheme for selecting a point with an intermediate index (see Figure 16 ).
[0462] The neighbor point set configuration unit 210090 may search for neighbor points of a point included in the LOD based on the LOD generated from the attribute information bitstream and register them as a neighbor point set. Figure 20 The neighbor point set configuration unit 20130 in.
[0463] The attribute information predictor 210100 may generate a predicted value of the attribute information. For example, the attribute information predictor may generate a predicted value of a point from a set of LOD and neighboring points, and generate a residual value of the predicted value. The attribute information predictor 210100 may perform the inverse process of the operation of the attribute information predictor 20140.
[0464] The residual attribute information inverse quantization processor 210110 may inversely quantize the residual attribute information (eg, the residual value of the attribute information). The residual attribute information inverse quantization processor 210110 may perform an inverse process of the operation of the residual attribute information quantization processor 20150.
[0465] The color inverse conversion processor 210120 may inversely convert the color information as the attribute information. The color inverse conversion processor 210120 may perform an inverse process of the operation of the color conversion processor 20100.
[0466] Figure 22 A bit stream structure of point cloud data according to an embodiment is shown.
[0467] The method / apparatus according to the embodiment may generate and obtain Figure 22 For example, a point cloud bitstream containing geometric information, attribute information and / or parameters including metadata can be obtained by Figure 1 The sending device 10000, the point cloud video encoder 10002, the receiving device 10004, the point cloud video decoder 10006, Figure 2 Encoding 20001 and decoding 20003 in Figure 4 The encoding process in Figure 11 The decoding process in Figure 12 The sending device in Figure 13 The receiving device in Figure 14 XR device 1430, Figure 20 The encoder in Figure 21 The decoder in generates (encodes) and receives (decodes).
[0468] Information of the implementation may be signaled.
[0469] The following abbreviations are used in the present disclosure: SPS (sequence parameter set); GPS (geometry parameter set); APS (attribute parameter set); TPS (patch parameter set); Geom (geometry bitstream (geometry bitstream = geometry slice header + geometry slice data)); Attr (attribute bitstream (=attribute block header + attribute block data)). Here, a block may be referred to as a block, etc.
[0470] According to the implementation scheme, the point cloud data may be taken as follows: Figure 19 The point cloud data may include a sequence parameter set (SPS), a geometry parameter set (GPS), an attribute parameter set (APS), and a tile parameter set (TPS), which may include signaling information according to an implementation. The point cloud data may include one or more geometries and / or attributes. The geometry and / or attributes in the point cloud data may be divided into one or more slices (or tiles / blocks). The geometry may have a structure of a geometry slice header and geometry slice data. For example, a TPS containing signaling information may include Tile(0)_tile_bounding_box_xyz0 and Tile(0)_tile_bounding_box_whd. The geometry may include geom_geom_parameter_set_id, geom_tile_id, geom_slice_id, geomBoxOrigin, geom_box_log2_scale, geom_max_node_size_log2, and geom_num_points.
[0471] Signaling information according to an embodiment may be added to SPS, GPS, APS, TPS, etc. and signaled.
[0472] According to an embodiment, the signaling information may be added to the TPS, the geometry (Geom) of each slice, or the attributes (Attr) of each slice and signaled.
[0473] The structure of point cloud data according to an embodiment can efficiently provide parameter sets, geometry, and attributes including signaling information in terms of encoding / decoding / data.
[0474] The point cloud data related to the point cloud data transmitting / receiving apparatus according to the embodiment may include at least one of a sequence parameter, a geometry parameter, an attribute parameter, a patch parameter, a geometry bit stream, or an attribute bit stream.
[0475] Hereinafter, the syntax of specific signaling information will be described with reference to the accompanying drawings. For reference, the name of the signaling information according to the implementation mode may be changed / extended within the intended meaning / functional scope of the signaling information. The fields of the signaling information may be distinguishably referred to as the first signaling information, the second signaling information, etc.
[0476] As described above, the point cloud data transmitting device (for example, referring to Figure 1 , Figure 11 , Figure 14 and Figure 20 The point cloud data transmitting device described herein may transmit the encoded point cloud data in the form of a bit stream. According to an embodiment, the bit stream may include one or more sub-bit streams.
[0477] Point cloud data transmitting device (for example, see Figure 1 , Figure 4 , Figure 12 and Figure 17 The point cloud data sending device described herein may divide the image of the point cloud data into one or more packets and send it via the network in consideration of errors in the transmission channel. The bitstream according to the embodiment may include one or more packets (e.g., network abstraction layer (NAL) units). Therefore, even when some packets are lost in a poor network environment, the point cloud data receiving device may use the remaining packets to restore the image. The point cloud data may be divided into one or more slices or one or more tiles and processed. The tiles and slices according to the embodiment are areas for segmenting a screen of point cloud data and processed by point cloud compression encoding. The point cloud data sending device may provide high-quality point cloud content by processing data corresponding to each area according to the importance of each divided area of the point cloud data. That is, the point cloud data sending device according to the embodiment may process data corresponding to areas important to the user with better compression efficiency and appropriate delay through point cloud compression encoding.
[0478] The image (or picture) of the point cloud content according to the embodiment is divided into basic processing units for point cloud compression encoding. The basic processing unit of the point cloud compression encoding according to the embodiment may include a coding tree unit (CTU) and a block, but is not limited to this example.
[0479] A slice according to an embodiment is an area including one or more basic processing units for point cloud compression encoding, and does not have a rectangular shape. A slice according to an embodiment contains data transmitted by grouping. A patch according to an embodiment is an area divided in a rectangular shape in an image, and includes one or more basic processing units for point cloud compression encoding. A slice according to an embodiment may be included in one or more patches. In addition, a patch according to an embodiment may be included in one or more slices.
[0480] The bitstream 3000 according to an embodiment may include signaling information and one or more slices, the signaling information including a sequence parameter set (SPS) for sequence level signaling, a geometry parameter set (GPS) for geometry information coding signaling, an attribute parameter set (APS) for attribute information coding signaling, and a patch parameter set (TPS) for patch level signaling.
[0481] The SPS according to an embodiment may be encoding information of the entire sequence such as a profile and a level, and may include comprehensive information about the entire file, such as a picture resolution and a video format.
[0482] According to an embodiment, a slice (e.g., Figure 22 The slice 0 of the GPS may include a slice header and slice data. The slice data may include a geometry bitstream (Geom00) and one or more attribute bitstreams (Attr00 and Attr10). The geometry bitstream may include a header (e.g., a geometry slice header) and a payload (e.g., geometry slice data). The header of the geometry bitstream according to the embodiment may include identification information (geom_geom_parameter_set_id) about the parameter set included in the GPS, a tile identifier (geom_tile id), a slice identifier (geom_slice_id), information about the origin of the geometry box (geomBoxOrigin), a geometry box scale (geom_box_log2_scale), a maximum geometry node size (geom_max_node_size_log2), the number of geometry points (geom_num_poins), and information about the data contained in the payload. The attribute bitstream may include a header (e.g., an attribute slice header or an attribute block header) and a payload (e.g., attribute slice data or attribute block data).
[0483] The method / apparatus according to an embodiment may add and signal the geometric quantization option information of each axis according to an embodiment to the SPS in the point cloud bitstream to the SPS. The geometric quantization option information of each axis according to an embodiment may be added to and signaled to the GPS. The SPS may maintain the scale value (quantization value) applied as a whole, and the GPS may set the scale value (quantization value) fine-tuned for each axis.
[0484] When quantization is performed differently for each tile, geometric quantization option information for each axis according to an embodiment may be added to the TPS and signaled. The geometric distribution may be different between tiles divided by region. Therefore, when the geometric quantization value (scale value) of each axis is set according to the distribution or the quantization value of each axis is readjusted according to the percentage of each axis to a quantization value, the geometric quantization of each axis may be performed differently for each tile.
[0485] The geometric quantization option information of each axis according to the embodiment may be added to the geometry header of each slice and signaled. The geometry distribution may be different between slices. Therefore, when the geometric quantization value (scale value) of each axis is set according to the distribution or the quantization value of each axis is readjusted according to the percentage of each axis to a quantization value, the geometric quantization of each axis may be performed differently for each slice.
[0486] Therefore, according to the quantization characteristics in the bitstream, the sending method / apparatus according to the implementation manner may include relevant parameter information (signaling information) in the bitstream, and may allow the receiving method / apparatus according to the implementation manner to statically / dynamically access the point cloud data.
[0487] The relevant information may be signaled to implement the implementation. The signaling information according to the implementation may be used at the transmitting side or the receiving side.
[0488] Geometry bitstream (Geom): Geometry bitstream = Geometry slice header + Geometry slice data
[0489] Attribute bitstream (Attr): Attribute bitstream = attribute block header + attribute block data.
[0490] The transmission method / apparatus according to the embodiment may add attribute point configuration related information of the spatial scalability function of the attribute encoding / decoding process to the SPS or APS and signal it.
[0491] The transmission method / apparatus according to the embodiment may add attribute point configuration related information of the spatial scalability function of the attribute encoding / decoding process to the TPS or Attr of each slice and notify it by signal.
[0492] According to an embodiment, tiles or slices are provided and point clouds are processed by segmenting them into regions. That is, point cloud data may be segmented into regions, wherein the segmentation unit may be tiles and / or slices.
[0493] When dividing the point cloud data into regions, the option of generating different sets of neighbor points for each region can be set, so that low complexity is obtained, but the reliability of the result is slightly lower, or high reliability is obtained, but the complexity is high. This can be set differently according to the processing capacity of the receiver.
[0494] When a point cloud is divided into tiles, different handling options for overlapping points can be applied to each tile.
[0495] When a point cloud is divided into slices, different processing options for overlapping points can be applied to each slice.
[0496] The attribute slice data may be changed and signaled according to options related to prediction mode processing.
[0497] Figures 23 to 26 Shown included in Figure 22 Detailed syntax of signaling information in the bit stream of point cloud data in. In order to support the operation according to the embodiment, the transmitting device and the receiving device according to the embodiment may provide a signaling scheme as described below.
[0498] Figure 23 A sequence parameter set according to an embodiment is shown.
[0499] Figure 23 Shown as Figure 22 The sequence parameter set included in the bitstream according to the embodiment is shown. Figure 23 The metadata, parameters or signaling information in Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Code 20001, Figure 2 Decoding 20003, Figure 4 The encoding process in Figure 11 The decoding process in Figure 12 The sending device in Figure 13 The receiving device in Figure 14 XR device 1430, Figure 20 The encoder in Generate, send and receive by decoders etc.
[0500] Information about the attribute point configuration of the spatial scalability function of the attribute encoding / decoding process can be added to the sequence parameter set and signaled. The syntax in can efficiently signal the attributes of supporting spatial scalability functions. The name of the signaling information can be understood within the scope of the meaning and function of the signaling information.
[0501] LOD_generation_direction indicates the LOD generation direction.
[0502] For example, when LOD_generation_direction is equal to 0, this indicates that LOD is generated from LOD0 to LODN (i.e., from the root node to the leaf node in the case of an octree). When LOD_generation_direction is equal to 1, this indicates that LOD is generated from LODN to LOD0 (i.e., from the leaf node to the root node in the case of an octree). The respective integer values of this information may vary according to implementation.
[0503] sampling_select_type indicates the point selection type of selecting one point from among points belonging to a node (including a merged node).
[0504] For example, when sampling_select_type is equal to 1, this indicates that the Nth point is selected. When sampling_select_type is equal to 2, this indicates that a point close to the center position of the octree parent node area is selected. When sampling_select_type is equal to 3, this indicates that a point with an intermediate index is selected. Each integer value of this information may change according to implementation.
[0505] sampling_isolated_node_min_number_of_points indicates the minimum number of points belonging to a node that can be defined as an isolated node.
[0506] sampling_isolated_node_merge_flag indicates whether to merge isolated nodes.
[0507] When sampling_isolated_node_merge_flag is 1, sampling_isolated_node_max_merge_range is signaled.
[0508] sampling_isolated_node_max_merge_range indicates the maximum merge range (number of consecutive nodes) for merging isolated nodes.
[0509] sampling_select_idx indicates the index of the point to be selected by the Nth point selection according to sampling_select_type. It may be set to 0 as a default value, which means selecting the first point.
[0510] When sampling_select_type is 1, sampling_select_idx is signaled.
[0511] profile_idc indicates the profile to which the bitstream conforms according to an embodiment. The bitstream shall not contain values of profile_idc other than those specified in the embodiments. Those values of profile_idc are reserved for future use by ISO / IEC.
[0512] When profile_compatibility_flags is equal to 1, it indicates that the bitstream complies with the profile indicated by profile_idc equal to j.
[0513] sps_num_attribute_sets indicates the number of coded attributes in the bitstream. The value of sps_num_attribute_sets may range from 0 to 63.
[0514] attribute_dimension[i] indicates the number of components of the i-th attribute.
[0515] attribute_instance_id[i] indicates the instance identifier of the i-th attribute.
[0516] A set of property parameters according to an embodiment is shown.
[0517] Shown as The sequence parameter set included in the bitstream according to the embodiment is shown. The metadata, parameters or signaling information in The sending device 10000, Point cloud video encoder 10002, The receiving device 10004 in Point cloud video decoder 10006, Code 20001, Decoding 20003, The encoding process in The decoding process in The sending device in The receiving device in XR device 1430, The encoder in Generate, send and receive by decoders etc.
[0518] aps_attr_parameter_set_id provides an identifier of the APS for reference by other syntax elements. The value of aps_attr_parameter_set_id may be in the range of 0 to 15, inclusive.
[0519] aps_seq_parameter_set_id specifies the value of sps_seq_parameter_set_id for the active SPS. The value of aps_seq_parameter_set_id can be in the range of 0 to 15 (inclusive).
[0520] If attr_coding_type is equal to 0 or 2, the lifting indication information (isLifting) is 1. Otherwise, isLifting is 0.
[0521] The promotion related signaling information is signaled as follows according to the promotion indication information.
[0522] lifting_num_pred_nearest_neighbours specifies the maximum number of nearest neighbors to be used for prediction. The value of lifting_num_pred_nearest_neighbours can be in the range of 1 to xx.
[0523] lifting_max_num_direct_predictors specifies the maximum number of predictors to be used for direct prediction. The value of lifting_max_num_direct_predictors may be in the range of 0 to lifting_num_pred_nearest_neighbours. The value of the variable MaxNumPredictors may be used in the decoding process as follows: MaxNumPredictors=lifting_max_num_direct_predicots+1.
[0524] For spatial_scalability_attribute_octree_level, LOD_generation_direction, sampling_select_type, sampling_isolated_node_min_number_of_points, sampling_isolated_node_merge_flag, sampling_isolated_node_max_merge_range, and sampling_select_idx according to an embodiment, refer to Description.
[0525] A patch parameter set according to an embodiment is shown.
[0526] Shown as The sequence parameter set included in the bitstream according to the embodiment is shown. The metadata, parameters or signaling information in The sending device 10000, Point cloud video encoder 10002, The receiving device 10004 in Point cloud video decoder 10006, Code 20001, Decoding 20003, The encoding process in The decoding process in The sending device in The receiving device in XR device 1430, The encoder in Generate, send and receive by decoders etc.
[0527] The transmitting apparatus or encoder according to the embodiment may signal the receiving apparatus or decoder according to the embodiment by adding property point configuration related information of spatial scalability for property encoding / decoding to a patch parameter set.
[0528] num_tiles specifies the number of tiles signaled to the bitstream. When not present, num_tiles is inferred to be 0.
[0529] tile_bounding_box_offset_x[i] indicates the x offset of the i-th tile in Cartesian coordinates. When not present, the value of tile_bounding_box_offset_x[0] is inferred to be sps_bounding_box_offset_x.
[0530] tile_bounding_box_offset_y[i] indicates the y offset of the i-th tile in Cartesian coordinates. When not present, the value of tile_bounding_box_offset_y[0] is inferred to be sps_bounding_box_offset_y.
[0531] For spatial_scalability_attribute_octree_level, LOD_generation_direction, sampling_select_type, sampling_isolated_node_min_number_of_points, sampling_isolated_node_merge_flag, sampling_isolated_node_max_merge_range, and sampling_select_idx according to an embodiment, refer to Description.
[0532] An attribute slice header according to an embodiment is shown.
[0533] Shown as The sequence parameter set included in the bitstream according to the embodiment is shown. The metadata, parameters or signaling information in The sending device 10000, Point cloud video encoder 10002, The receiving device 10004 in Point cloud video decoder 10006, Code 20001, Decoding 20003, The encoding process in The decoding process in The sending device in The receiving device in XR device 1430, The encoder in Generate, send and receive by decoders etc.
[0534] The transmitting device or the encoder according to the embodiment may add attribute point configuration related information for the spatial scalability function in the attribute encoding / decoding process to the attribute slice header and signal it.
[0535] ash_attr_parameter_set_id specifies the value of aps_attr_parameter_set_id of the active APS.
[0536] ash_attr_sps_attr_idx specifies the attribute set in the active SPS. The value of abh_attr_sps_attr_idx shall be in the range of 0 to sps_num_attribute_sets in the active SPS.
[0537] ash_attr_geom_slice_id specifies the value of gsh_slice_id of the active geometry slice header.
[0538] The PCC encoding method, PCC decoding method, and signaling method according to the above-described embodiments may provide the following effects.
[0539] Spatial scalability is a function of using narrower bandwidth (e.g., thumbnails) when point cloud data is dense, or sending and processing lower resolution point clouds to reduce the amount of computation on the decoder side.
[0540] An embodiment provides a method for improving attribute compression efficiency by changing a scheme of sampling points for each LOD based on an octree, so that an encoder / decoder of geometry-based point cloud compression (G-PCC) for 3D point cloud data compression provides a spatial scalability function so that the most suitable points of each LOD are configured as nodes.
[0541] The proposed method of reorganizing points for each octree node through Morton code can reduce the encoding / decoding execution time by reducing the computation amount of encoding / decoding constituting the octree.
[0542] Considering the distance weight based on the center position value of the bounding box of the parent node applied for spatial scalability through the proposed sampling scheme of each node, as proposed in this paper, points close to the center position value or points with intermediate indexes can be selected to reduce the error value of the distance weight to enhance the attribute compression efficiency.
[0543] Therefore, an embodiment may provide an attribute compression method for providing a spatial scalability function of an encoder / decoder of a geometry-based point cloud compression (G-PCC) for compressing 3D point cloud data and may increase attribute compression efficiency.
[0544] The PCC encoder and / or PCC decoder according to the embodiment may have a spatial scalability effect by providing a method of changing a point sampling technique for each LOD based on an octree.
[0545] Next, a geometry / attribute compression method for compressing 3D point cloud data according to an embodiment for supporting spatial scalability of an encoder / decoder of G-PCC will be described.
[0546] The embodiment proposes a method of changing a method of setting a geometric point position value based on an octree and a point sampling method of each LOD to provide a spatial scalability function in an attribute encoding / decoding process of a G-PCC.
[0547] The method according to the embodiment includes the following contents.
[0548] The present invention proposes 1) a method for calculating geometric positions in partial recovery for supporting spatial scalability, 2) a method for sampling reference points for each node, and 3) a signaling method for supporting the above methods.
[0549] Reference Embodiments relate to a method for compressing properties to support spatial scalability of G-PCC for 3D point cloud data compression. When point cloud data is dense, it may be desirable to use a narrower bandwidth (e.g., thumbnails) or to send and process lower resolution point clouds to reduce computational complexity at the decoder side. This functionality is referred to as spatial scalability.
[0550] The transmission / reception method / apparatus according to the embodiment may provide a spatial scalability function through a process of encoding or decoding only occupied bits up to a depth level selected by adjusting a depth level of an octree.
[0551] Attribute encoding / decoding according to an embodiment is shown.
[0552] The geometrical position is shown to provide The method of spatial scalability function is shown.
[0553] The method in can be The encoder in The decoder and / or the corresponding sending / receiving device according to the implementation manner are executed.
[0554] In order to provide spatial scalability functionality even for attribute encoding / decoding, the octree structure may need to be modified and applied. The method of generating LOD starting from a selected depth level of the octree, the method of calculating neighbor nodes, and the method of calculating weights may need to be changed to provide spatial scalability in attribute encoding / decoding.
[0555] The embodiment proposes setting the geometric position value in the geometric part decoding. The embodiment also provides a technique for changing the sampling points for each LOD based on an octree and configuring the most suitable point as a node for each LOD to provide spatial scalability in attribute encoding / decoding to improve attribute compression efficiency.
[0556] The embodiment provides a method for changing the point sampling type of each LOD based on an octree to provide a spatial scalability function in the attribute encoding / decoding process of G-PCC.
[0557] Embodiments may set representative position values of an octree for providing spatial scalability in geometry decoding. In order to provide spatial scalability for attribute encoding / decoding, an octree structure may be applied in geometry encoding / decoding. Embodiments provide a method for reorganizing points belonging to a node starting from a selected depth level of an octree and a method for each LOD sampling point.
[0558] The method of reorganizing points and the method for each LOD sampling point are performed in PCC attribute encoding / decoding of a PCC decoder / encoder, and can be applied when a predictive transform technique or a lifting transform technique is used in the attribute encoding / decoding process.
[0559]
[0560] When decoding a point in a point cloud, it is necessary to perform decoding only up to a certain octree depth and recover the position value of the point. A function supporting this operation may be spatial scalability. When decoding is partially performed at an intermediate depth of the octree, the node closest to the center position of the bounding box of each octree node may be sampled (selected) as the position of the point representing each octree node. In this case, when an attribute is partially decoded, Morton order, LOD configuration, and neighbor node search may be implemented based on the position set to the intermediate position of the node.
[0561] S270010 Sorting based on the Morton order
[0562] Points in the point cloud may be converted to Morton codes based on x, y, and z position values and may be sorted based on the Morton code values.
[0563] S270020 Set the attribute part encoding / decoding area
[0564] In attribute encoding / decoding, the depth level of the octree can be set, and points to which attribute encoding is to be applied can be constructed and encoded / decoded starting from the set octree depth level, as in geometry encoding / decoding (see Fig.17 ).
[0565] S270030 Use Morton codes to reorganize points belonging to nodes
[0566] The method of reorganizing the points sorted by Morton code for each octree node can be referred to Fig.16 The following pseudo code describes the.
[0567] nodeSizeLog2: Octree depth level
[0568] PX_mortoncode: PX Morton code
[0569] end_range_mc: the Morton code range of the node
[0570] Int maxNumOfPoints=pow(8,nodeSizeLog2+1);
[0571] For(int I=1;I <indexCount;++i){
[0572] if(Pi_mortoncode <end_range_mc){
[0573] …
[0574] }else{
[0575] end_range_mc=Pi_mortoncode+
[0576] (maxNumOfPoints-(Pi_mortoncode%maxNumOfPoints));
[0577] }
[0578] }
[0579] The Morton code range of a node can be calculated based on the Morton code of the point, the octree depth level, and the maximum number of points.
[0580] like Fig.18 As shown, end_range_mc can be estimated from the Morton code of the point. When the Morton code value of the next point is within the range of end_range_mc, the points may belong to the same node. When the Morton code value of the next point is outside the range, this may mean that the node has changed to a new node. In this case, end_range_mc can be estimated based on the Morton code of the point and applied to subsequent points. In this way, points belonging to an octree node can be reorganized from points sorted by Morton code.
[0581] S270030 LOD sampling at octree nodes
[0582] Sampling can be performed by selecting points from among the points belonging to a node (including merged nodes). Attribute compression may be affected by the selected points. The best reference points are selected and reorganized into a reference region (retained list), and a set of neighbor nodes is constructed from the reference region. Therefore, compression efficiency can be increased.
[0583] When the position of the octree node of the geometry is changed to the center point, the error generated at the parent node can be reduced by selecting a position close to the center position value as the point selected in the attribute encoding. Therefore, the LOD sampling at the octree node can adopt the method of selecting a position close to the center of the parent node.
[0584] There are many ways to find a location close to the center. The location can be calculated to find the location correctly. To reduce the time required to find the location, the location on the octree node can be referenced. Various options given below can be used. Figures 28 to 30 Describes sampling: 1) Select the first node at the LOD, then select the last node at the next consecutive upper LOD (see Figures 28 to 30 ); 2) select a point close to the center of the octree parent node region; 3) select a point with an intermediate index; and / or 4) a combination thereof.
[0585] A method of receiving point cloud data according to an embodiment may include: receiving a bit stream including the point cloud data; and decoding the point cloud data.
[0586] Point cloud data according to an embodiment may include geometric data and attribute data.
[0587] According to an embodiment, decoding of point cloud data may include decoding of geometry data and decoding of attribute data.
[0588] The geometric data decoding according to the embodiment may include: generating an octree structure for the geometric data; and sampling (selecting) positions of points representing respective nodes included in the octree structure as nodes closest to the center positions of bounding boxes of the respective nodes.
[0589] According to an embodiment, attribute data decoding may include: changing position values of points in point cloud data into Morton codes based on geometric data and attribute data; sorting the points based on the Morton codes; and decoding the points starting from a depth level of an octree structure including the points.
[0590] The attribute data decoding according to the embodiment may further include reorganizing points belonging to each node based on the Morton code and the range value of the Morton code.
[0591] According to an embodiment, attribute data decoding may include: generating an octree structure for geometric data; generating a level of detail (LOD) based on the octree structure; and sampling (selecting) a node belonging to the LOD from the LOD. That is, sampling a point included in the node.
[0592] Fig.28 Bounding boxes and nodes are shown according to an embodiment.
[0593] Fig.28 express Fig. 27 S270010 in.
[0594] Eight nodes may be included in the bounding box. When sorted in the order of the Morton code, points having values 0 to 7 as Morton code values may be arranged in sequence.
[0595] Fig.29 The matching between various nodes and locations according to an embodiment is shown.
[0596] Fig.29 Show when the point Fig.28 The operation that matches each node to its position when sorted in the order of the Morton code. Fig.29 express Fig. 27 S270030 in.
[0597] The reference node (or reserved node) represents the node (point) selected during sampling. For example, Fig.29 Nodes (or points) 29000, 29040 and 29050 in Fig.30 A node (or point) in can be a reference node.
[0598] As shown in the figure, when the first node 29000 (LOD N) is selected, the node closest to the center position 29030 of the bounding box 29020 of the parent node 29010 (LOD N-1) of the first node may be the last node 29040. When the last node 29040 is selected at LOD N-1, the node closest to the center position of its parent node at LOD N-2 may be the first node 29050.
[0599] Therefore, in LOD sampling, when LOD is even, the first node may be selected, and when LOD is odd, the last node may be selected. Alternatively, when LOD is even, the last node may be selected, and when LOD is odd, the first node may be selected. The encoder may signal the configuration information about the above selection scheme to the decoder.
[0600] That is, the encoder and / or decoder according to the embodiment provides various schemes to select the node closest to the center position on the parent node in the child node when performing LOD sampling at the octree node. The embodiment includes a scheme using even indexes and odd indexes or a scheme for selecting the node closest to the center position of the bounding box. For example, the node closest to the center position may be selected based on the distance, density, and attributes of the point. In addition, when selecting a point close to the center position, if the selected point belongs to a 2×2×2 node, the node close to the center may be sampled. Otherwise, odd / even nodes (i.e., the first / last node) may be sampled.
[0601] Fig.30 FIG. 1 illustrates selecting a reference node in an octree structure according to an embodiment.
[0602] Fig.30 Shows the octree structure Fig.28 and Fig.29 LOD sampling ( Fig. 27 S270030 in the.
[0603] In the octree structure according to the embodiment, the nodes may be distributed sequentially from the root node to the leaf node, and the next node belonging to the node may be sorted in the order of the Morton code. The octree structure may include N LODs, and the transmission / reception method / device according to the embodiment may perform partial encoding and partial decoding starting from a specific depth level in the octree structure.
[0604] In the octree structure, the points belonging to LOD N are sorted in the order of the Morton code, and the reference node is the node sampled at each LOD. When LOD N (where N can be called the index of the LOD) is an even LOD, the first point can be selected from the points. When LOD N-1 is an odd LOD, the last point can be selected from the points. As the operation proceeds from a specific node to the parent node in the octree structure, the nodes can be selected in the order of the first node, the last node, the first node, the last node, etc. This scheme can increase the gain of sampling.
[0605] For example, for point 30000 belonging to LOD N, when LOD N is assumed to be an even LOD, the first node 30010 among the points belonging to the node sorted in the order of the Morton code can be selected (sampled) as the reference node (or reserved node) 30010 of each parent node.
[0606] For the point 30020 belonging to LOD N-1, N-1 may be regarded as an odd number, and the last node 30030 among the points belonging to the node sorted in the order of the Morton code may be selected (sampled) as a reference node for each parent node.
[0607] According to an embodiment, sampling may be performed by sampling the first node at the level where LOD sampling starts, sampling the last node at the next level, and sampling the first node at the next level. Alternatively, sampling may be performed by sampling the last node at the level where LOD sampling starts, sampling the first node at the next level, and sampling the last node at the next level.
[0608] This operation can be expressed as pseudo code as follows.
[0609] endIndex=0
[0610] assignedPointCount=0
[0611] for(lod=minGeomNodeSizeLog2; unprocessedPointCount>0; lod++){
[0612] startIndex=assignedPointCount
[0613] nonAssignedPointCount=0
[0614] samplingFromLastFlag = lod & 1 / / If LOD is even, the last bit is assigned 0, if LOD is odd, it is assigned 1. / /
[0615] for(i=0;i <unprocessedPointCount;i++){
[0616] currVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i]]>>(3×(lod+1))
[0617] if(samplingFromLastFlag==0){
[0618] if (i == 0) / / This means that the LOD is even, and the first point is sampled as the reference node.
[0619] nonAssignedPointIndexes[nonAssignedPointCount++]= unprocessedPointIndexes[i]
[0620] else{
[0621] prevVoxelIndex=McodeUnsorted[unprocessedPointIndexes[i-1]]>>(3×(lod +1))
[0622] if(currVoxelIndex>prevVoxelIndex)
[0623] nonAssignedPointIndexes[nonAssignedPointCount++]= unprocessedPointIndexes[i]
[0624] (Here, sampled points are assigned to NonassignedPointsIndexes. Non-sampled points are assigned to assignedPointIndexes. By setting NonassignedPointsIndexes and assignedPointIndexes separately, the method / apparatus according to the embodiment can reorganize the LOD.)
[0625] else
[0626] assignedPointIndexes[assignedPointCount++]=unprocessedPointIndexes[i]
[0627] }
[0628] }else{ / / This means that the LOD is even and the last point is sampled as the reference node.
[0629] if(i==(unprocessedPointCount–1))
[0630] nonAssignedPointIndexes[nonAssignedPointCount++]= unprocessedPointIndexes[i]
[0631] else{
[0632] nextVoxelIndex = McodeUnsorted[unprocessedPointIndexes[i + 1]]>>(3×(lod + 1))
[0633] if (currVoxelIndex < nextVoxelIndex)
[0634] nonAssignedPointIndexes[nonAssignedPointCount++]= unprocessedPointIndexes[i]
[0635] (Here, NonassignedPointsIndexes represents the sampled points.)
[0636] else
[0637] assignedPointIndexes[assignedPointCount++] = unprocessedPointIndexes[i]
[0638] }
[0639] }
[0640] }
[0641] endIndex = assignedPointCount
[0642] if (startIndex != endIndex) {
[0643] numOfPointInSkipped = PointNumInSlice – PointCount
[0644] if ((endIndex – startIndex) > (startIndex + numOfPointInSkipped)) {
[0645] for (loop = 0; loop < lod - minGeomNodeSizeLog2; loop++) {
[0646] computeNearestNeighbours(
[0647] PointCount - unprocessedPointCountPerLevelOfDetail[loop],
[0648] PointCount - unprocessedPointCountPerLevelOfDetail[loop + 1],
[0649] loop + minGeomNodeSizeLog2, assignedPointIndexes,
[0650] McodeUnsorted, nonAssignedPointCount,
[0651] nonAssignedPointIndexes)
[0652] }
[0653] }
[0654] }
[0655] computeNearestNeighbours(
[0656] startIndex, endIndex,
[0657] lod, assignedPointIndexes,
[0658] McodeUnsorted, nonAssignedPointCount,
[0659] nonAssignedPointIndexes)
[0660] unprocessedPointCountPerLevelOfDetail[lod + 1] = nonAssignedPointCount
[0661] unprocessedPointCount = nonAssignedPointCount
[0662] unprocessedPointIndexes = nonAssignedPointIndexes
[0663] }
[0664] Then, apply the following process:
[0665] for (i = 0; i < PointCount; i++)
[0666] indexes[PointCount - 1 – i] = assignedPointIndexes[i]
[0667] for(lod=0; lod <LevelDetailCount;lod++)
[0668] pointCountPerLevelOfDetail[lod]=
[0669] unprocessedPointCountPerLevelOfDetail[LevelDetailCount-1-lod]
[0670] / / The value of samplingFromLastFlag indicates whether sampling can be started from the first node (or left) or the last node (or right) at the starting LOD.
[0671] According to an embodiment, if signaling information such as a lifting LOD sampling related flag (lifting_lod_regular_sampling_enabled_flag) included in the bitstream indicates an LOD sampling scheme, the LOD may be reorganized according to the octree nodes based on the geometry data (if lifting_scalability_enabled_flag is equal to 1, the level of detail generation process reorganizes the points into a set of refinement levels (R_l)_(l=0…L-1) according to the octree nodes based on the geometry. Depending on the value of the sampling related flag (samplingFromLastFlag), the first point in the node (i.e., samplingFromLastFlag is equal to 0) or the last point in the node (i.e., samplingFromLastFlag is equal to 1) is sampled).
[0672] Next, a scheme of selecting a point close to the center of an octree parent node area and a scheme of selecting a point with an intermediate index according to an embodiment will be described.
[0673] Select a point close to the center of the octree parent node region:
[0674] The sending / receiving method / device according to the embodiment can select the point closest to the center position of the bounding box of the parent node when selecting a point from the points belonging to the node (including the merged node). Selecting a point closer to the center position of the parent node can reduce the error when selecting neighbor points and calculating the weight of neighbor points.
[0675] When selecting neighbor points, the distance between points is calculated, and points closer in distance can be selected as neighbor points. In order to provide spatial scalability function, the position value of a point can be replaced by the value of the center position of the node to which the point belongs (the node can vary according to the level of LOD), and then the distance can be calculated. Therefore, in order to find the neighbor point of point Pi, the position value of Pi can be changed to the center position of the parent node according to LOD1, and the position of the point Ri in the neighbor point candidate list (reserved list) R also changes according to LOD1. Then, the distance of the point can be calculated. In other words, the distance between the center positions of nodes at the same depth level can be calculated, so that the neighbor node is set as the neighbor point. From the distance calculated in this way, the weight of the neighbor point can be obtained, 1 / distance=Pi. Since the distance is calculated by changing the position of the point to the center position of the parent node, the calculation contains errors. As the depth level of the node decreases, that is, the bounding box of the node becomes larger, the error can increase.
[0676] In order to reduce the error, a point close to the center of the parent node area can be selected as the reference point. The error can be reduced according to the position value of the reference point, and the reduced error can have a good effect on the compression efficiency.
[0677] Select the point with the middle index:
[0678] When selecting a point from among the points belonging to a node (including a merged node), the transmission / reception method / device according to the embodiment may select a point with an intermediate index from among the points belonging to the node (including a merged node). Since the probability that the node with the intermediate index is close to the center position of the bounding box of the parent node is high (in many cases), and selecting a point by the intermediate index does not require a lot of additional calculations, the probability of finding a point close to the center position with a small amount of calculation can be high.
[0679] The method of receiving point cloud data according to an embodiment may include sampling (selecting) a first point or a last point in a node belonging to the LOD based on the LOD.
[0680] Fig.31 A PCC data encoder according to an embodiment is shown.
[0681] Fig.31 The encoder in corresponds to Fig. 20 The point cloud data encoder and / or the corresponding device according to the embodiment. Fig.31 For the operation of each component in Fig. 20 Description.
[0682] The LOD configuration unit 31000 may perform LOD configuration and sampling according to an embodiment. The LOD configuration unit 31000 may use Morton code to reorganize points belonging to a node, merge isolated nodes, perform LOD sampling in an octree node according to an embodiment, and notify relevant information with a signal.
[0683] The LOD configuration unit 31000 may set the depth level of the octree to support spatial scalability in attribute encoding / decoding, and configure and encode / decode points to which attribute encoding is to be applied starting from the set octree depth level (see Fig.17 and Fig.30 ).
[0684] A transmitting device or an encoder according to an embodiment may signal an octree depth level as signaling information to a receiving device or a decoder according to an embodiment, and the octree depth level may be changed according to its purpose or the computational performance of the decoder.
[0685] The LOD construction unit 31000 may perform point reorganization for each octree node by calculating the Morton code range of each octree node at the points sorted by the Morton code (see Fig.18 and Fig. 27 ).
[0686] When the LOD component 31000 selects a point from among the points belonging to a node, a selection scheme of the operation may be signaled to the decoder. The selection scheme may include a scheme of selecting the first node at an even LOD and the last node at an odd LOD, a scheme of selecting a point close to the center of the octree parent node region, and a scheme of selecting a point with an intermediate index (see Figures 28 to 30 ).
[0687] In a scheme where the first node is selected at an LOD and then the last node is selected at the next consecutive LOD, it may be determined whether the first node or the last node is selected at an even LOD. This determination may be signaled to the decoder (see Figures 28 to 30 ).
[0688] The residual attribute information quantization processor 31010 can generate a predicted value of the attribute data, generate a residual value between the predicted value and the attribute value, and quantize the residual value.
[0689] When the prediction mode and residual information separation flag is true, the residual attribute information quantization processor 31010 can calculate the residual between the predicted value according to the prediction mode and the value of the point, and encode its quantized value together with the prediction mode.
[0690] When the prediction mode and residual information separation flag is false, the residual attribute information quantization processor 31010 may calculate the residual between the prediction value and the point value according to the prediction mode, and encode only its quantized value.
[0691] Fig.32 A PCC data decoder according to an embodiment is shown.
[0692] Fig.32 The decoder in corresponds to Fig.21 The point cloud data decoder and / or the corresponding device according to the embodiment. Fig.32 For the operation of each component in Fig.21 Description.
[0693] The octree reconstructor 32000 reconstructs the octree from the received geometry information bitstream (geometry data). For the spatial scalability function, the octree reconstructor 32000 may construct geometry information only at a specific level of the octree. In this case, the position value may be changed to the value of the center of the bounding box.
[0694] The octree reconstructor 32000 may construct and decode points to which attribute decoding is to be applied for partial geometry decoding starting from the restored octree depth level.
[0695] The LOD configuration unit 32010 may use Morton code to reorganize points belonging to a node, merge isolated nodes, and perform LOD sampling in an octree node according to an implementation.
[0696] The LOD configuration unit 32010 may construct and decode points to which attribute decoding is to be applied starting from the octree depth level restored for partial attribute decoding.
[0697] The LOD configuration unit 32010 may perform point reorganization for each octree node by calculating a Morton code range of each octree node from points sorted by Morton code.
[0698] The LOD configuration unit 32010 may select a point from among the points belonging to a node (including a merged node) based on a scheme for selecting a restored point. The scheme may include a scheme for selecting the first node at the LOD and then selecting the last node at the next consecutive upper LOD, a scheme for selecting a point close to the center position of the octree parent node area, and a scheme for selecting a point with an intermediate index.
[0699] In a scheme where a first node is selected at a LOD and then a last node is selected at a next consecutive upper LOD, the LOD configuration unit 32010 may configure the LOD by selecting the first node and the last node according to the restored selection type.
[0700] Figure 33 to Figure 36 Show Fig. 22 Detailed syntax of signaling information contained in the bit stream of point cloud data in. In order to support the operation according to the embodiment, the transmitting device and the receiving device according to the embodiment may provide a signaling scheme as described below.
[0701] The transmitting device and the receiving device according to the embodiment can be added to the SPS ( Fig.33 ) or APS( Fig.34) to signal information related to the configuration of attribute points of the spatial scalability function.
[0702] The transmitting device and the receiving device according to the embodiment can be added to the TPS ( Fig.35 ) or the Attr( Fig.36 ) is used to signal information related to the configuration of the attribute points of the spatial scalability function. Due to the signaling information according to the embodiment, the geometry / attributes for supporting the spatial scalability function can be efficiently signaled.
[0703] Fig.33 A sequence parameter set (SPS) according to an embodiment is shown.
[0704] Fig.33 Shown as Fig. 22 The sequence parameter set included in the bitstream according to the embodiment is shown. Fig.33 The metadata, parameters or signaling information in Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Code 20001 in Figure 2 Decoding 20003, Figure 4 The encoding process in Fig.11 The decoding process in Fig.12 The sending device in Fig.13 The receiving device in Fig.14 XR device 1430, Fig. 20 The encoder in Fig.21 Generate, send and receive by decoders etc.
[0705] Information on the configuration of the attribute points of the spatial scalability function may be added to the sequence parameter set and signaled.
[0706] spatial_scalability_octree_level indicates the partially restored octree level.
[0707] spatial_scalability_attribute_octree_level indicates the partially restored octree level. That is, it indicates the partial octree level applied to the attribute. Attributes can have a different restored octree level than the geometry.
[0708] LOD_generation_direction indicates the LOD generation direction.
[0709] For example, when LOD_generation_direction is equal to 0, this indicates that LOD is generated from LOD0 to LODN (i.e., from the root node to the leaf node in the case of an octree). When LOD_generation_direction is equal to 1, this indicates that LOD is generated from LODN to LOD0 (i.e., from the leaf node to the root node in the case of an octree). The integer value indicating the direction may vary depending on the implementation.
[0710] sampling_select_type indicates the point selection type when selecting a point from among the points belonging to a node (including a merge node). For example, when sampling_select_type is equal to 1, this indicates that the first node is selected at an even LOD and the last node is selected at an odd LOD. When sampling_select_type is equal to 2, this indicates that a point close to the center position of the octree parent node area is selected. When sampling_select_type is equal to 3, this indicates that a point with an intermediate index is selected. The integer value indicating the type may vary depending on the implementation.
[0711] sampling_first_node_flag indicates whether the first node or the last node is selected at the starting LOD.
[0712] When sampling_select_type == 1, sampling_first_node_flag may be provided as signaling information.
[0713] For other fields, refer to Fig.23 Description.
[0714] Fig.34 An attribute parameter set (APS) according to an embodiment is shown.
[0715] Fig.34 Shown as Fig. 22 The sequence parameter set included in the bitstream according to the embodiment is shown. Fig.34 The metadata, parameters or signaling information in Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Code 20001, Figure 2 Decoding 20003, Figure 4 The encoding process in Fig.11 The decoding process in Fig.12 The sending device in Fig.13 The receiving device in Fig.14 XR device 1430, Fig. 20 The encoder in Fig.21 Generate, send and receive by decoders etc.
[0716] Information about the configuration of the attribute points of the spatial scalability function may be added to the attribute parameter set and signaled.
[0717] For spatial_scalability_attribute_octree_level, LOD_generation_direction and sampling_select_type, sampling_first_node_flag, refer to Fig.33 Description.
[0718] For other fields, refer to Fig.24 Description.
[0719] Fig.35 A tile parameter set (TPS) according to an embodiment is shown.
[0720] Fig.35 Shown as Fig. 22 The sequence parameter set included in the bitstream according to the embodiment is shown. Fig.35 The metadata, parameters or signaling information in Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Code 20001, Figure 2 Decoding 20003, Figure 4 The encoding process in Fig.11 The decoding process in Fig.12 The sending device in Fig.13 The receiving device in Fig.14 XR device 1430, Fig. 20 The encoder in Fig.21 Generate, send and receive by decoders etc.
[0721] Information about the configuration of the attribute points of the spatial scalability function may be added to the tile parameter set and signaled.
[0722] For spatial_scalability_attribute_octree_level, LOD_generation_direction and sampling_select_type, sampling_first_node_flag, refer to Fig.33 Description.
[0723] For other fields, refer to Fig.25 Description.
[0724] Fig.36 An attribute slice header (ASH) according to an embodiment is shown.
[0725] Fig.36 Shown as Fig. 22 The sequence parameter set included in the bitstream according to the embodiment is shown. Fig.36 The metadata, parameters or signaling information in Figure 1 The sending device 10000, Figure 1 Point cloud video encoder 10002, Figure 1 The receiving device 10004 in Figure 1 Point cloud video decoder 10006, Figure 2 Code 20001 in Figure 2 Decoding 20003, Figure 4 The encoding process in Fig.11 The decoding process in Fig.12 The sending device in Fig.13 The receiving device in Fig.14 XR device 1430, Fig. 20 The encoder in Fig.21 Generate, send and receive by decoders etc.
[0726] Information on the configuration of the attribute points of the spatial scalability function may be added to the attribute slice header and signaled.
[0727] For spatial_scalability_attribute_octree_level, LOD_generation_direction and sampling_select_type, sampling_first_node_flag, refer to Fig.33 Description.
[0728] For other fields, refer to Fig.26 Description.
[0729] Fig.37 A method of transmitting point cloud data according to an embodiment is shown.
[0730] S37000: The method for transmitting point cloud data according to an embodiment may include encoding the point cloud data. The encoding operation according to an embodiment may be performed with Figure 1 The sending device 1000, Figure 1 Point cloud video encoder 10002, Figure 2 Code 20001 in Figure 4 The encoder in Fig.12 The encoder in Fig.14 The XR device 1430 in Fig.15 The partial decoding corresponds to the partial encoding in Figures 16 to 19 The encoding process in Fig. 20 The encoder in Figure 22 to Figure 26 Point cloud data bitstream processing in Figures 27 to 30 The encoding process in Fig.31 The encoder in Figure 33 to Figure 36 The operations of point cloud data bit stream processing correspond to or are combined.
[0731] S37010: The method for sending point cloud data according to an embodiment may further include sending a bit stream containing point cloud data. The sending operation according to an embodiment may be performed with Figure 1 The transmitter 10003 in Figure 2 Transmission 20002 and Fig.12 The operations of the transmission processor 12012 correspond to or are combined.
[0732] Fig.38 A method of receiving point cloud data according to an embodiment is shown.
[0733] S38000: The method for receiving point cloud data according to an embodiment may include receiving a bit stream containing point cloud data. The receiving operation according to an embodiment may be performed with Figure 1 Receiver 10005 in Figure 2 Transmission 20002 and Fig.13 The operations of the receiver 13000 correspond to or are combined.
[0734] S38010: The method for receiving point cloud data according to an embodiment may further include decoding the point cloud data. The decoding operation according to an embodiment may be performed with Figure 1 Point cloud video decoder 10006, Figure 2 Decoding 20003, Fig.10 The decoder in Fig.11 The decoder in Fig.13 The decoder in Fig.14 XR device 1430, Fig.15 Partial decoding in Figures 16 to 19 The decoding process in Fig.21 The decoder in Figure 22 to Figure 26 Point cloud data bit stream processing and decoding Figures 27 to 30 The decoding process in Fig.32 The decoder in Figure 33 to Figure 36 The operations corresponding to or combined with point cloud data bit stream processing and the like.
[0735] The PCC encoding method, PCC decoding method, and signaling method according to the above-described embodiments may provide the following effects.
[0736] Spatial scalability is a function of using narrower bandwidth (e.g., thumbnails) when point cloud data is dense, or sending and processing lower resolution point clouds to reduce the amount of computation on the decoder side.
[0737] An embodiment provides a method for improving attribute compression efficiency by changing a scheme for restoring the center position value of a bounding box of each node based on an octree as a position value of a point cloud and sampling points for each LOD, so that an encoder / decoder of geometry-based point cloud compression (G-PCC) for 3D point cloud data compression provides a spatial scalability function so that the most appropriate point of each LOD is configured as a node.
[0738] The proposed method of reorganizing points for each octree node through Morton code can reduce the encoding / decoding execution time by reducing the computation amount of encoding / decoding constituting the octree.
[0739] Considering the distance weight based on the center position value of the bounding box of the parent node applied for spatial scalability through the proposed sampling scheme of each node, as proposed in this paper, points close to the center position value or points with intermediate indexes can be selected to reduce the error value of the distance weight to enhance the attribute compression efficiency.
[0740] Therefore, an embodiment may provide an attribute compression method for providing a spatial scalability function of an encoder / decoder of a geometry-based point cloud compression (G-PCC) for compressing 3D point cloud data and may increase attribute compression efficiency.
[0741] A PCC encoder and / or PCC decoder according to an embodiment may provide a spatial scalability function by providing a scheme for restoring a point position of a geometry to a center position value of a node based on an octree and changing a point sampling scheme for each LOD.
[0742] By proposing a method for calculating geometric positions in partial recovery for supporting spatial scalability, a method for sampling reference points for each node, and a signaling method supporting the above method, a spatial scalability effect on attribute encoding / decoding of G-PCC can be provided.
[0743] In this document, "A / B" is interpreted as A and / or B.
[0744] The embodiments have been described from the perspective of the method and / or the apparatus, and the descriptions of the method and the apparatus may be applied to supplement each other.
[0745] Although the drawings are described separately for simplicity, new embodiments can be designed by merging the embodiments shown in the various figures. Designing a recording medium that can be read by a computer and records a program for executing the above-mentioned embodiments according to the needs of those skilled in the art also falls within the scope of the attached claims and their equivalents. The apparatus and method according to the embodiment may not be limited to the configuration and method of the above-mentioned embodiment. Various modifications may be made to the embodiment by selectively combining all or some of the embodiments. Although preferred embodiments are described with reference to the drawings, it will be understood by those skilled in the art that various modifications and changes may be made to the embodiment without departing from the spirit or scope of the present disclosure described in the attached claims. These modifications should not be understood separately from the technical ideas or viewpoints of the embodiments.
[0746] Various elements of the device of the embodiment can be implemented by hardware, software, firmware or a combination thereof. Various elements in the embodiment can be implemented by a single chip (e.g., a single hardware circuit). According to the embodiment, the components according to the embodiment can be implemented as separate chips respectively. According to the embodiment, at least one or more components of the device according to the embodiment may include one or more processors capable of executing one or more programs. One or more programs may execute any one or more operations / methods according to the embodiment or include instructions for executing them. The executable instructions for executing the method / operation of the device according to the embodiment may be stored in a non-transitory CRM or other computer program product configured to be executed by one or more processors, or may be stored in a temporary CRM or other computer program product configured to be executed by one or more processors. In addition, the memory according to the embodiment can be used as not only covering volatile memory (e.g., RAM), but also covering the concepts of non-volatile memory, flash memory and PROM. In addition, it can also be implemented in the form of a carrier (e.g., transmission via the Internet). In addition, the processor-readable recording medium can be distributed to a computer system connected via a network so that the processor-readable code can be stored and executed in a distributed manner.
[0747] In the present specification, the terms " / " and "," should be interpreted as indicating "and / or". For example, the expression "A / B" may mean "A and / or B". In addition, "A, B" may mean "A and / or B". In addition, "A / B / C" may mean "at least one of A, B and / or C". In addition, "A / B / C" may mean "at least one of A, B and / or C". In addition, in the present specification, the term "or" should be interpreted as indicating "and / or". For example, the expression "A or B" may mean 1) only A, 2) only B, or 3) both A and B. In other words, the term "or" used in this document should be interpreted as indicating "in addition or alternatively".
[0748] Terms such as first and second can be used to describe various elements of the embodiments. However, various components according to the embodiments should not be limited by the above terms. These terms are only used to distinguish one element from another element. For example, a first user input signal can be referred to as a second user input signal. Similarly, a second user input signal can be referred to as a first user input signal. The use of these terms should not be interpreted without departing from the scope of various embodiments. Both the first user input signal and the second user input signal are user input signals, but unless the context clearly stipulates otherwise, they do not mean the same user input signal.
[0749] The terms used to describe the embodiments are used to describe specific embodiments and are not intended to limit the embodiments. As used in the description of the embodiments and the claims, unless the context clearly stipulates otherwise, the singular form includes plural referents. The expression "and / or" is used to include all possible combinations of terms. Terms such as "including" or "having" are intended to indicate the presence of graphics, quantities, steps, elements and / or components, and should be understood as not excluding the possibility of additional presence of graphics, quantities, steps, elements and / or components. As used herein, conditional expressions such as "if" and "when" are not limited to optional situations, and are intended to be interpreted as performing related operations when specific conditions are met, or interpreting related definitions according to specific conditions.
[0750] The operations according to the embodiments described in this specification may be performed by a transmitting / receiving device including a memory and / or a processor according to the embodiments. The memory may store a program for processing / controlling the operations according to the embodiments, and the processor may control the various operations described in this specification. The processor may be referred to as a controller, etc. In an embodiment, the operations may be performed by firmware, software, and / or a combination thereof. The firmware, software, and / or a combination thereof may be stored in a processor or a memory.
[0751] Invention Mode
[0752] As described above, the relevant contents are described in the best mode for carrying out the embodiment.
[0753] Industrial Applicability
[0754] As described above, the embodiments may be applied in whole or in part to a point cloud data transmitting / receiving device and system.
[0755] It will be apparent to those skilled in the art that various changes or modifications may be made to the embodiments within the scope of the embodiments.
[0756] Thus, it is intended that the embodiments cover the modifications and variations of this disclosure provided they come within the scope of the appended claims and their equivalents.
Claims
1. A method comprising the following steps: receiving a bitstream comprising point cloud data, The point cloud data includes geometric data and attribute data; Decoding the geometric data; as well as decoding the attribute data, The geometric data includes an octree, Generate the level of detail LOD based on the octree, Sampling the points in the nodes of the LOD, where the point with the lowest point index in an even level is assigned to the output of that level, and Among them, the point with the maximum point index in an odd-numbered level is assigned to the output of that level.
2. The method according to claim 1, in, The step of decoding the geometric data comprises: generating an octree for the geometric data, The position of the point representing each node of the octree is set to the node closest to the center position of the bounding box for each node.
3. The method according to claim 1, in, The step of decoding the attribute data comprises: changing the position value of the point of the point cloud data into a Morton code based on the geometric data and the attribute data, arranging the points based on the Morton code, The point is decoded from a depth level of the octree that includes the point.
4. The method according to claim 3, in, The step of decoding the attribute data comprises: Points included in respective nodes of an octree are arranged based on the Morton code and a range value of the Morton code.
5. A method comprising the following steps: Encode the geometric data of the point cloud data; encoding the attribute data of the point cloud data, The geometric data includes an octree, Generate the level of detail LOD based on the octree, Sampling points in the nodes of the LOD; generating a bitstream including the point cloud data, where the point with the lowest point index in an even level is assigned to the output of that level, and Among them, the point with the maximum point index in an odd-numbered level is assigned to the output of that level.
6. The method according to claim 5, in, The step of encoding the geometric data comprises: generating an octree for the geometric data, The position of the point representing each node of the octree is set to the node closest to the center position of the bounding box for each node.
7. The method according to claim 5, in, The step of encoding the attribute data comprises: changing the position values of the points of the point cloud data into Morton codes based on the geometric data and the attribute data, arranging the points based on the Morton codes, The point is encoded from a depth level of the octree that includes the point.
8. The method according to claim 7, in, The step of encoding the attribute data comprises: Points included in respective nodes of an octree are arranged based on the Morton code and a range value of the Morton code.
9. A method comprising the following steps: Encode the geometric data of the point cloud data; encoding the attribute data of the point cloud data, The geometric data includes an octree, Generate the level of detail LOD based on the octree, Sampling points in the nodes of the LOD; generating a bitstream comprising the encoded geometry data and the encoded attribute data; as well as sending the bit stream, wherein the point of the lowest point index in an even-numbered level is assigned to the output of that level, and wherein the point of the maximum point index in an odd-numbered level is assigned to the output of that level.
10. A computer-readable storage medium storing a bit stream generated by the method according to claim 5.
