Method for updating a time evolving scan
Patent Information
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-07-04
- Publication Date
- 2026-05-27
Smart Images
Figure EP2024068821_23012025_PF_FP_ABST
Abstract
Description
[0001] METHOD FOR UPDATING A TIME EVOLVING SCAN
[0002] 1. Technical Field
[0003] The present principles generally relate to the domain of Augmented or Extended Reality applications. The present document is also understood in the context of generating a scene description for an evolving Extended Reality scene and of updating the evolving Extended Reality scene when changes are detected, for example for a rendering of the evolving Extended Reality scene on end-user devices such as mobile devices or Head-Mounted Displays (HMD).
[0004] 2. Background
[0005] The present section is intended to introduce the reader to various aspects of art, which may be related to various aspects of the present principles that are described and / or claimed below. This discussion is believed to be helpful in providing the reader with background information to facilitate a better understanding of the various aspects of the present principles. Accordingly, it should be understood that these statements are to be read in this light, and not as admissions of prior art.
[0006] In Augmented Reality (AR) or Extended Reality (XR) experiences, virtual content is seamlessly inserted into the user real environment that is captured by using cameras or video- see-through devices. A description of elements of the user real environment is required for different tasks like the positioning of virtual objects attached to AR anchors, consistent managing of collisions between virtual and real objects, or consistent rendering of virtual and real objects including occlusion and lighting / shadowing aspects. Some existing AR frameworks are able to capture, compute, store and load real environment data for AR experiences. Real environment data are computed from embedded-sensor raw data. An AR device may have several embedded sensors to scan the real environment, such as color camera(s) and / or Light Detection and Ranging (LiDAR). Generated raw data may be point clouds, depth maps and / or pictures. An Inertial Measurement Unit (IMU) is also required to estimate the current pose (i.e. location and orientation in the 3D space of the real environment) of the AR device when acquiring these data. Based on these sensor raw data, a representation of the real environment is computed, and the resulting real environment data may have various formats. A mesh representation may be sufficient for coherent collision handling and lighting. However, a semantic representation (e.g. “desk”, “laptop”, “screen”, “floor”, “ceiling”, “wall”) may be required for the definition of advanced anchoring and / or interaction. A mesh segmentation is required for detecting and managing individual real object. So, a description of the real environment requires a scanning step and a semantic segmentation step. These steps are complicated and time consuming.
[0007] When a scan of a real environment is loaded, some elements of the real environment may have been moved since the scan has been performed. Re-scanning the objects that have moved (local scan) and starting a new segmentation is a performance and time expansive method. Moreover, the new scan may be performed from a different location and orientation than the first scan and a complex matching step is required in addition to the semantic segmentation. There is a lack of a technique to allow the user to locally and optionally repair a first scan described in an AR / XR scene description file.
[0008] 3. Summary
[0009] The following presents a simplified summary of the present principles to provide a basic understanding of some aspects of the present principles. This summary is not an extensive overview of the present principles. It is not intended to identify key or critical elements of the present principles. The following summary merely presents some aspects of the present principles in a simplified form as a prelude to the more detailed description provided below.
[0010] The present principles relate to a method for generating a scene description of extended reality scene. The method comprises segmenting a scan of the extended reality scene in mesh objects. Then, for each object, when the object is fully visible in the scan, a first status and a third status are attributed to the object. Otherwise, when is object is partially visible in the scan, a second status is attributed to the object. The method checks whether the object is completable and, if so, replaces the partial mesh by a complete mesh for the object and attributes the third status to the object. Then, a scene description is generated as a node tree and the third status is attributed to nodes corresponding to objects having the third status.
[0011] The present principles also relate to a device comprising a memory associated with a processor configured for implementing the method above. 4. Brief Description of Drawings
[0012] The present disclosure will be better understood, and other specific features and advantages will emerge upon reading the following description, the description making reference to the annexed drawings wherein:
[0013] - Figure 1 diagrammatically illustrates the required steps for generating a description of an AR scene representative of a real environment;
[0014] - Figure 2 illustrates a scene description formatted as a tree;
[0015] - Figure 3 shows an example architecture of a device 30 which may be configured to implement a method for generating a scene description for an evolving Extended Reality scene and of updating the evolving Extended Reality scene when changes are detected according to the present principles;
[0016] - Figure 4 shows an example of an embodiment of the syntax of a stream when the data are transmitted over a packet-based transmission protocol;
[0017] - Figure 5 diagrammatically illustrates the initial step that allows to insert the "canRepair" flag in the scene description file and the build of the scene description file according to the present principles;
[0018] - Figure 6 diagrammatically illustrate a runtime processing model according to the present principles.
[0019] 5. Detailed description of embodiments
[0020] The present principles will be described more fully hereinafter with reference to the accompanying figures, in which examples of the present principles are shown. The present principles may, however, be embodied in many alternate forms and should not be construed as limited to the examples set forth herein. Accordingly, while the present principles are susceptible to various modifications and alternative forms, specific examples thereof are shown by way of examples in the drawings and will herein be described in detail. It should be understood, however, that there is no intent to limit the present principles to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present principles as defined by the claims.
[0021] The terminology used herein is for the purpose of describing particular examples only and is not intended to be limiting of the present principles. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises", "comprising," "includes" and / or "including" when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. Moreover, when an element is referred to as being "responsive" or "connected" to another element, it can be directly responsive or connected to the other element, or intervening elements may be present. In contrast, when an element is referred to as being "directly responsive" or "directly connected" to other element, there are no intervening elements present. As used herein the term "and / or" includes any and all combinations of one or more of the associated listed items and may be abbreviated as" / ".
[0022] It will be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element without departing from the teachings of the present principles.
[0023] Although some of the diagrams include arrows on communication paths to show a primary direction of communication, it is to be understood that communication may occur in the opposite direction to the depicted arrows.
[0024] Some examples are described with regard to block diagrams and operational flowcharts in which each block represents a circuit element, module, or portion of code which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that in other implementations, the function(s) noted in the blocks may occur out of the order noted. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending on the functionality involved.
[0025] Reference herein to “in accordance with an example” or “in an example” means that a particular feature, structure, or characteristic described in connection with the example can be included in at least one implementation of the present principles. The appearances of the phrase in accordance with an example” or “in an example” in various places in the specification are not necessarily all referring to the same example, nor are separate or alternative examples necessarily mutually exclusive of other examples. Reference numerals appearing in the claims are by way of illustration only and shall have no limiting effect on the scope of the claims. While not explicitly described, the present examples and variants may be employed in any combination or sub-combination.
[0026] An AR device may have several embedded sensors to scan the real environment, such as color camera(s) and Light Detection and Ranging (LiDAR). Generated raw data may be point clouds, depth maps and / or pictures. An Inertial Measurement Unit (IMU) is also required to estimate the current pose (i.e. location and orientation in the 3D space of the real environment) of the AR device when acquiring these data. Based on these sensor raw data, a representation of the real environment is computed, and the resulting real environment data may have various formats. A mesh representation may be sufficient for coherent collision handling and lighting. However, a semantic representation (e.g. “desk”, “laptop”, “screen”, “floor”, “ceiling”, “wall”) may be required for the definition of advanced anchoring and / or interaction. A mesh segmentation is required for detecting and managing individual real object. So, a description of the real environment requires a scanning step and a semantic segmentation step. These steps are complicated and time consuming.
[0027] Figure 1 illustrates the required steps for generating a description of an AR scene representative of a real environment. For example, Microsoft Mixed Reality framework have been developed for the HoloLens 2 device. It is composed of a spatial computing module generating a mesh representation of the real environment as depict by picture 11 of Figure 1 and of a scene understanding module from Mixed Reality Toolkit (MRTK) version 2.7 based on OpenXR, detecting and labeling planar surfaces for the placement of virtual content. On a fourth-generation iPad Pro running iPad OS 13.4 or later, Apple’s ARKit uses a LiDAR Scanner to create a mesh representation of the user real environment. This mesh is further segmented and multiple anchors, called ARMeshAnchor, and is assigned to the resulting set of segmented meshes. A semantic labeling is performed for the real objects that ARKit can identify such as ceiling, door, floor, seat, table, wall and window labels as illustrated by picture 12 of Figure 1. The Meta / Oculus framework has ben developed for Meta Quest 2 and Meta Quest Pro devices. The scene understanding system provides a scene model, that is a representation of the user real environment. Floor, ceiling, wall face, desk, couch, door frame and window_frame labels are currently supported. A Scene Model is generated by the Scene Capture system flow that lets users walk around and capture their scene. In every case, after capturing the scan of a scene, a segmentation operation can be launched. The semantic segmentation of meshes has been a research topic explored for several years. A framework is characterized by the number of labels available in its database. The higher this number, the more difficult it will be to perform the segmentation operation in real time. Deep learning has gained significant success in 3D segmentation, a survey of the different framework is given. For example, ScanNet provides online benchmark for 3D semantic segmentation evaluation on the test set, it’s a reference for mesh segmentation framework. The computation of the real environment data may either be done locally in the AR device or remotely in a computing server. A processes to complete segmented meshes may, for example, replace segment objects by their 3D models in the scene or replace by a full model provided that a full model has been scanned (for example, chairs in a room).
[0028] So, when an incomplete scan information fails identify an object, but a complete model of the object can be obtained from a database of known objects, this complete model is used to complete the scan. It implies that, if the object was rotated in the real world between sessions and that a previously unscanned side of the object is actually displayed, that object still can be identify since the complete model of the object can be obtained from the objects database, which stores correct features from all sides. According to the present principles, three steps are involved in this process. First, efficiently identifying the areas of the scene that are consistent with the previously stored local scene model, i.e. the areas that don’t require any modification to the model. Second, efficiently identifying discrepancy areas that can be repaired, and do those repairs either automatically, or with user assistance. Third, falling back to local rescanning (with segmentation / object identification) of any areas remaining after first and second steps. Such a repair action may be possible when the volume of the bounding box of a given object represents less than a threshold, for example 90 or 8 or 1 percent of the volume of the bounding box of the 3D scene.
[0029] A scene is composed of real (for example a scan of a room) and virtual assets. Real assets are useful to increase AR / XR experience. For example, the scan of the room allows management of collisions between virtual objects and real objects. This can also be used for rendering purpose as for instance, to not render real objects in the case of a see-through XR device or to ensure coherent lighting between real and virtual objects including shadow management. A scene can be advantageously represented by a graph composed of nodes corresponding to the real and virtual objects. Solutions to indicate the nature of every node of a scene description exist, for example through flag “MPEG node nature” in MPEG standards. A segmented scene is composed of a set of real nodes that improve the AR / XR experience. A scan of the scene can be stored and reloaded later for a new AR / XR experience, avoiding a new scan operation.
[0030] When a first scan of a real environment is loaded, some elements of the real environment may have been moved since the first scan has been performed. According to the present principles, when a segmented mesh of a real environment is available from first scan and semantic segmentation processes, a possibility is displayed to the user to modify the scan locally and optionally, providing a handle to a fallback mesh. Two main steps are performed according to the present principles: processing of a segmented mesh to set up a flag “canRepair” and inserting the flag “canRepair” in a scene description file. Flag “canRepair” is used by the rendering device at runtime. When the scanned scene would have locally evolved, this flag will be used to launch repair actions. So, interactions with the segmented mesh are enhanced and simplified. According to the present principles, when possible, the model is updated on the server side.
[0031] Figure 2 illustrates a scene description formatted as a tree. After the first scan and the associated semantic segmentation process, the result of segmentation is checked according to the present principles. Nodes of the scene description tree are labelled according to three statuses: Complete (The segmentation of the object corresponds to a completed model), Partial (only part of the object is extracted) and Unknow (the object is not identified; nothing will be done). A flag in relation with the previous status of the node in a scene description file (for example glTF file) is added to follow changes. The use of a scene description file allows distributing an AR / XR scene to several users from a single server. This also allows to reload a model of a scanned scene without performing the scan and the semantic segmentation again.
[0032] After a segmentation process, a status ‘Complete’, ‘Partial’ or ‘Unknown’ is associated with nodes of the scene description tree as illustrated in Figure 2. Status ‘Complete’ means that the object has been fully scanned and recognized. Status ‘Partial’ stands for areas of the object have not totally been scanned but the object is recognized. For example, a chair under a table is not fully scanned but a label ‘chair’ can be associated with this object mesh. This operation may be performed automatically by the segmentation framework or driven by a user. When a status ‘Complete’ is associated with a node, a repair action is possible, a corresponding displaying is rendered. Different factors, like the size of the object are also taken into account at this step. For example, a repair action is possible when the volume of the bounding box of a given object represents less than a threshold, for example 10 or 50 or 2 percent of the volume of the bounding box of the 3D scene.
[0033] Figure 5 diagrammatically illustrates the initial step that allows to insert the "canRepair" flag in the scene description file and the build of the scene description file. At a step 51, the scene is scanned, and a semantic segmentation is performed at a step 52. Then a segmentation status is associated with nodes at step 53. Status “Complete” and status “Partial” are in relation with a threshold (recognition rate). They are determined by comparison with an existing model in a database. If the determined status is ‘Partial’, then a check for completion step 54 is performed. If this check step is successful, then, the object associated with the node is replaced by the completed object at step 55. When the determined status is ‘Complete’ or when the check for completion is successful, the canRepair flag is associated with the object at step 56. Then another segmented object segmented at step 52 is considered until every object has been considered.
[0034] The use of a scene description file allows an export to a server to share this file. As part of a scene description, parameters can be added in a standard as shown in the following example based on MPEG-I Scene Description. MPEG-I Scene Description is a framework using the Khronos glTF extension mechanism to support additional scene description features. The semantic of the MPEG node nature is provided in the following table.
[0035] According to the present principles, the semantic of MPEG node nature is extended as follows:
[0036] Where usage ‘M’ stands for ‘Mandatory’ and ‘O’ for ‘Optional’.
[0037] As described herein above, moving an object may reveal holes in the mesh. A fallback mesh may have been provided to work around this problem. In the example scene description below, meshes 0 and 4 are impacted. In row “meshFallback”, a value -1 indicates that there is no fallback. In this example, there is a fallback only for mesh 0.
[0038]
[0039]
[0040] Figure 6 diagrammatically illustrate a runtime processing model according to the present principles. At the beginning of an AR / XR experience, the application loads the scene description file at a step 60. If a segmented scan of a real scene with an extension including the flag “canRepair” is read in the scene description at a step 61, then the user can apply a repair action 62 when the scan and the scene are not fully aligned.
[0041] When the AR / XR application loads a first (previous) scan, if the mesh of an object is not in agreement with the real object, a graphic interface may be displayed to allow the user to move the mesh to superimpose it to the real object. An update of the pose of the mesh can be sent to the server if an uplink exists, in this case only a TRS (Translation / Rotation / Scale values) is transmitted. This operation may have impacts on other parts of the mesh. In this case, a fallback operation is implemented. For example, when a chair is placed on the floor, there will be holes in the mesh of the floor at the point of contact with the legs of the chair. So, if the mesh of the chair is displaced, holes will be visible. In the case of the floor, the mesh may be replaced by a 3D model of plan. For other cases, like the relighting of the real scene by a virtual light, it may be necessary to add a texture to the repaired mesh. When the user identifies that there is a difference between the loaded model and the real scene, he can launch the process of a repair action. In this embodiment, this action is performed based on visual observation. In another embodiment, an automatic comparison mode is considered by matching a picture captured from a known point of view with the textured mesh. In a third embodiment, an automatic comparison of the meshes of the first scan with the meshes of a second (new) scan is performed.
[0042] In an embodiment, the repair process is used when an object does no longer belong to the scene. In this case, the mesh is deleted. In another embodiment, the fallback can be a geometric primitive (sphere, cube, disc... ) used as a model of a real object.
[0043] Figure 3 shows an example architecture of a device 30 which may be configured to implement a method for generating a scene description for an evolving Extended Reality scene and of updating the evolving Extended Reality scene when changes are detected according to the present principles. The device may implement the methods described in relation to Figure 5 and to Figure 6. Alternatively, each circuit of the encoder and / or the decoder may be a device according to the architecture of Figure 3, linked together, for instance, via their bus 31 and / or via I / O interface 36. Device 30 comprises following elements that are linked together by data and address bus 31 :
[0044] - a microprocessor 32 (or CPU), which is, for example, a DSP (or Digital Signal Processor);
[0045] - a ROM (or Read Only Memory) 33;
[0046] - a RAM (or Random Access Memory) 34;
[0047] - a storage interface 35;
[0048] - an I / O interface 36 for reception of data to transmit, from an application; and
[0049] - a power supply, e.g. a battery.
[0050] In accordance with an example, the power supply is external to the device. In each of mentioned memory, the word « register » used in the specification may correspond to area of small capacity (some bits) or to very large area (e.g. a whole program or large amount of received or decoded data). The ROM 33 comprises at least a program and parameters. The ROM 33 may store algorithms and instructions to perform techniques in accordance with present principles. When switched on, the CPU 32 uploads the program in the RAM and executes the corresponding instructions.
[0051] The RAM 34 comprises, in a register, the program executed by the CPU 32 and uploaded after switch-on of the device 30, input data in a register, intermediate data in different states of the method in a register, and other variables used for the execution of the method in a register.
[0052] The implementations described herein may be implemented in, for example, a method or a process, an apparatus, a computer program product, a data stream, or a signal. Even if only discussed in the context of a single form of implementation (for example, discussed only as a method or a device), the implementation of features discussed may also be implemented in other forms (for example a program). An apparatus may be implemented in, for example, appropriate hardware, software, and firmware. The methods may be implemented in, for example, an apparatus such as, for example, a processor, which refers to processing devices in general, including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device. Processors also include communication devices, such as, for example, computers, cell phones, portable / personal digital assistants ("PDAs"), and other devices that facilitate communication of information between end-users.
[0053] In accordance with examples, the device 30 belongs to a set comprising:
[0054] - a mobile device;
[0055] - a communication device;
[0056] - a game device;
[0057] - a tablet (or tablet computer);
[0058] - a laptop;
[0059] - a still picture camera;
[0060] - a video camera;
[0061] - an encoding chip;
[0062] - a server (e.g. a broadcast server, a video-on-demand server or a web server).
[0063] Figure 4 shows an example of an embodiment of the syntax of a stream when the data are transmitted over a packet-based transmission protocol. Figure 4 shows an example structure 4 of a stream encoding a scene description of an Extended Reality scene according to the present principle. The structure consists in a container which organizes the stream in independent elements of syntax. The structure may comprise a header part 41 which is a set of data common to every syntax element of the stream. For example, the header part comprises some of metadata about syntax elements, describing the nature and the role of each of them. The structure comprises a payload comprising an element of syntax 42 and at least one element of syntax 43. Element of syntax 42 comprises the scene description itself, for example organized a node tree. Element of syntax 43 is a part of the payload of the data stream and may comprise data referenced by the nodes of the scene description. There may be an element of syntax 43 for different types of data, for instance one for the nodes, one for the meshes, one for the textures, etc.
[0064] The implementations described herein may be implemented in, for example, a method or a process, an apparatus, a computer program product, a data stream, or a signal. Even if only discussed in the context of a single form of implementation (for example, discussed only as a method or a device), the implementation of features discussed may also be implemented in other forms (for example a program). An apparatus may be implemented in, for example, appropriate hardware, software, and firmware. The methods may be implemented in, for example, an apparatus such as, for example, a processor, which refers to processing devices in general, including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device. Processors also include communication devices, such as, for example, Smartphones, tablets, computers, mobile phones, portable / personal digital assistants ("PDAs"), and other devices that facilitate communication of information between end-users.
[0065] Implementations of the various processes and features described herein may be embodied in a variety of different equipment or applications, particularly, for example, equipment or applications associated with data encoding, data decoding, view generation, texture processing, and other processing of images and related texture information and / or depth information. Examples of such equipment include an encoder, a decoder, a post-processor processing output from a decoder, a pre-processor providing input to an encoder, a video coder, a video decoder, a video codec, a web server, a set-top box, a laptop, a personal computer, a cell phone, a PDA, and other communication devices. As should be clear, the equipment may be mobile and even installed in a mobile vehicle.
[0066] Additionally, the methods may be implemented by instructions being performed by a processor, and such instructions (and / or data values produced by an implementation) may be stored on a processor-readable medium such as, for example, an integrated circuit, a software carrier or other storage device such as, for example, a hard disk, a compact diskette (“CD”), an optical disc (such as, for example, a DVD, often referred to as a digital versatile disc or a digital video disc), a random access memory (“RAM”), or a read-only memory (“ROM”). The instructions may form an application program tangibly embodied on a processor-readable medium. Instructions may be, for example, in hardware, firmware, software, or a combination. Instructions may be found in, for example, an operating system, a separate application, or a combination of the two. A processor may be characterized, therefore, as, for example, both a device configured to carry out a process and a device that includes a processor-readable medium (such as a storage device) having instructions for carrying out a process. Further, a processor-readable medium may store, in addition to or in lieu of instructions, data values produced by an implementation.
[0067] As will be evident to one of skill in the art, implementations may produce a variety of signals formatted to carry information that may be, for example, stored or transmitted. The information may include, for example, instructions for performing a method, or data produced by one of the described implementations. For example, a signal may be formatted to carry as data the rules for writing or reading the syntax of a described embodiment, or to carry as data the actual syntax-values written by a described embodiment. Such a signal may be formatted, for example, as an electromagnetic wave (for example, using a radio frequency portion of spectrum) or as a baseband signal. The formatting may include, for example, encoding a data stream and modulating a carrier with the encoded data stream. The information that the signal carries may be, for example, analog or digital information. The signal may be transmitted over a variety of different wired or wireless links, as is known. The signal may be stored on a processor-readable medium.
[0068] A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made. For example, elements of different implementations may be combined, supplemented, modified, or removed to produce other implementations. Additionally, one of ordinary skill will understand that other structures and processes may be substituted for those disclosed and the resulting implementations will perform at least substantially the same function(s), in at least substantially the same way(s), to achieve at least substantially the same result(s) as the implementations disclosed. Accordingly, these and other implementations are contemplated by this application.
Claims
CLAIMS1. A method for generating a scene description of extended reality scene, the method comprising:- obtaining a scan of the extended reality scene segmented in mesh objects;- for each object,• when the scan of the object is fully visible in the scan, attributing to the object, and a flag indicating that the object can be repaired;• when the scan of the object is partially visible in the scan, checking whether the object is completable and, if so, replace the partial mesh by a complete mesh for the object and attribute the flag indicating that the object can be repaired to the object; and- generating a scene description as a node tree and attributing the flag to nodes corresponding to objects having the flag.
2. The method of claim 1, wherein a semantic segmentation of the scan is performed to determine whether an object is fully visible.
3. The method of claim 2, wherein the determination whether an object is fully visible is performed by comparison to a model in a database according to the semantic segmentation.
4. The method of one of claims 1 to 3, wherein a first status (complete) is attributed to a node when the scan of the object is fully visible, a second status (partial) is attributed to the node when the scan of the object is partially visible, and a third status (unknown) is attributed to the node in other cases.
5. The method of one of claims 1 to 4, wherein nodes of the node tree comprise information indicating whether the node corresponds to a real object.
6. The method of one of claims 1 to 5, wherein the scene description is encoded in a data stream.
7. A device for generating a scene description of extended reality scene, the device comprising a processor configured:- segmenting a scan of the extended reality scene in mesh objects;- for each object,• when the object is fully visible in the scan, attributing to the object, and a flag indicating that the object can be repaired;• when is object is partially visible in the scan, checking whether the object is completable and, if so, replace the partial mesh by a complete mesh for the object and attribute the flag indicating that the object can be repaired to the object; and- generating a scene description as a node tree and attributing the flag to nodes corresponding to objects having the flag.
8. The device of claim 7, wherein the processor is configured for performing a semantic segmentation of the scan is performed to determine whether an object is fully visible.
9. The device of claim 8, wherein the determination whether an object is fully visible is performed by comparison to a model in a database according to the semantic segmentation.
10. The device of one of claims 7 to 9, wherein a first status (complete) is attributed to a node when the scan of the object is fully visible, a second status (partial) is attributed to the node when the scan of the object is partially visible, and a third status (unknown) is attributed to the node in other cases.
11. The device of one of claims 7 to 10, wherein nodes of the node tree comprise information indicating whether the node corresponds to a real object.
12. The device of one of claims 7 to 11, wherein the scene description is encoded in a data stream.