Method and apparatus for compression and encoding of 3d dynamic point cloud
By determining predictors and using context-based encoding for 3D point clouds, the method enhances compression efficiency by leveraging temporal redundancy, improving compression ratios while maintaining performance in coding and decoding times.
Patent Information
- Application Number
- GB2022015237
- Authority / Receiving Office
- GB · GB
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-10-14
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2042-10-14
AI Technical Summary
Current 3D point cloud compression techniques, such as G-PCC V1, independently compress frames without exploiting temporal redundancy, leading to suboptimal compression ratios, which the proposed INTER prediction tools aim to address by leveraging temporal redundancy between consecutive frames.
The method involves determining a set of predictors in a reference frame for predictive encoding of 3D points, selecting a predictor based on criteria such as collocated points, and encoding the index using a context-based probability distribution, with improvements including the use of two arithmetic contexts for encoding the INTER predictor index to enhance compression efficiency.
This approach improves compression ratios and maintains coding and decoding times, effectively addressing the limitations of independent frame compression by exploiting temporal redundancy in 3D dynamic point clouds.
Smart Images

Figure 00000001_0000 
Figure 00000002_0000 
Figure 00000003_0000
Abstract
Description
FIELD OF THE INVENTION The present disclosure concerns a method and a device for compression and encoding of 3D point cloud. It concerns more particularly 3D dynamic point cloud. A 3D dynamic point cloud is a temporal sequence of 3D point clouds. Each 3D point cloud comprising a variable number of 3D points with variable 3D positions. BACKGROUND OF INVENTION In particular, the 3D point cloud can be a set of 3D points captured by a LIDAR located on the top of a car. For example, the LIDAR can be a rotating LIDAR containing different elevation lasers with a horizontal rotation along the vertical dimension. For compressing 3D points, a standard called G-PCC V1 (ISO / IEC FDIS 23090-9:2022(E)) has emerged which proposes to encode 3D points based on INTRA prediction. However, recent improvements over the standard are proposed in a MPEG document called Technology Under Consideration (“ISO / IEC JTC 1 / SC 29 / WG 7 N00281”, called herein TUC). In particular, the TUC document proposes to use INTER prediction, prediction between consecutive frames, a frame being in the present case a 3D point cloud associated with a same capture instant, by analogy to the video compression techniques where the frame is a picture associated with a capture instant in a video sequence. INTER prediction tools have been recently accepted as potential adopted compression tools for the future version of G-PCC. The compression (or encoding, both terms being used in this document referring to the same techniques) techniques proposed in these two documents are based on predictive encoding where the encoding of a particular point is based on the identification of a predictor, and the encoding of a residual constituted by the difference between the current point and the 3D coordinates of a previously encoded point (the predictor). As the encoding may be destructive, the residual is actually computed as the difference of the current point and the decoded version of the predictor. INTRA modes correspond to modes where the predictor is chosen in the current frame among the previously encoded / decoded points, meaning the current 3D point cloud. INTER modes correspond to modes where the predictor is chosen in a reference frame, meaning a previously encoded / decoded 3D point cloud. Improvements in term of coding efficiency may be advantageous for both INTRA and INTER modes. SUMMARY OF THE INVENTION The present invention has been devised to address one or more of the foregoing concerns. It concerns different improvements over techniques defined in the standard, the TUC propositions and adopted tools for future version of G-PCC. According to a first aspect of the invention it is proposed a method of encoding a 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, in a bitstream, wherein the method comprises for encoding a current 3D point in a current frame of 3D points: - determining a set of predictors in a reference frame that may be used for predictive encoding of the current 3D point; - determining an index of a selected predictor in the set of predictors, the selected predictor being used for encoding the current 3D point; - determining a class in a set of classes to which the current 3D point belongs based on a criterion; and - encoding, using a context based on a probability distribution of previously encoded indexes in the determined class, the determined index. In an embodiment, the criterion is based on the existence of one or more collocated point in the reference frame corresponding respectively to one or more previously encoded point in the current frame. In an embodiment, the criterion is based on the existence of a collocated point in the reference frame corresponding to a previously encoded point in the current frame. In an embodiment, the set of predictors point comprises only two predictors, and wherein the index is a Boolean value indicating which predictor is the selected predictor. According to another aspect of the invention it is proposed a method of decoding a bitstream comprising an encoded 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, wherein the method comprises for decoding a current 3D point in a current frame of 3D points: - determining a set of predictors points in a reference frame that may be used for predictive encoding of the current 3D point; - determining a class in a set of classes to which the current 3D point belongs based on a criterion; and - decoding an index of a selected predictor in the set of predictors, from the bitstream, using a context based on a probability distribution of previously encoded indexes in the determined class. According to another aspect of the invention it is proposed a device for encoding a 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, in a bitstream, wherein the device comprises for encoding a current 3D point in a current frame of 3D points a processor configured for: - determining a set of predictors in a reference frame that may be used for predictive encoding of the current 3D point; - determining an index of a selected predictor in the set of predictors, the selected predictor being used for encoding the current 3D point; - determining a class in a set of classes to which the current 3D point belongs based on a criterion; and - encoding, using a context based on a probability distribution of previously encoded indexes in the determined class, the determined index. According to another aspect of the invention it is proposed a device for decoding a bitstream comprising an encoded 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, wherein the device comprises for decoding a current 3D point in a current frame of 3D points a processor configured for: - determining a set of predictors points in a reference frame that may be used for predictive encoding of the current 3D point; - determining a class in a set of classes to which the current 3D point belongs based on a criterion; and - decoding an index of a selected predictor in the set of predictors, from the bitstream, using a context based on a probability distribution of previously encoded indexes in the determined class. According to another aspect of the invention it is proposed a computer program product for a programmable apparatus, the computer program product comprising a sequence of instructions for implementing a method according to the invention, when loaded into and executed by the programmable apparatus. According to another aspect of the invention it is proposed a computer-readable storage medium storing instructions of a computer program for implementing a method according to the invention. According to another aspect of the invention it is proposed a computer program which upon execution causes the method of the invention to be performed. At least parts of the methods according to the invention may be computer implemented. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a "circuit", "module" or "system". Furthermore, the present invention may take the form of a computer program product embodied in any tangible medium of expression having computer usable program code embodied in the medium. Since the present invention can be implemented in software, the present invention can be embodied as computer readable code for provision to a programmable apparatus on any suitable carrier medium. A tangible, non-transitory carrier medium may comprise a storage medium such as a floppy disk, a CD-ROM, a hard disk drive, a magnetic tape device or a solid state memory device and the like. A transient carrier medium may include a signal such as an electrical signal, an electronic signal, an optical signal, an acoustic signal, a magnetic signal or an electromagnetic signal, e.g. a microwave or RF signal. BRIEF DESCRIPTION OF THE DRAWINGS Embodiments of the invention will now be described, by way of example only, and with reference to the following drawings in which: Figure 1 illustrates an example of LIDAR system; Figure 2 illustrates different compression tools proposed in this standard or different new adopted tools for the future version of the standard; Figure 3 illustrates an example of the main steps of a general encoding method for a point cloud and the construction of the associated bitstream; Figure 4 illustrates the main steps of an example of decoding method for decoding a bitstream generated by the encoder; Figure 5 illustrates examples of points obtained by a rotating lidar; Figure 6 illustrates an example of tree construction for a rotating lidar; Figure 7 illustrates the main step of an example of the compression algorithm used for encoding the 3D geometry of the input points; Figure 8 illustrates a schematic version of an example of encoding process for choosing between INTER and INTRA compression according to embodiments of the invention; Figure 9 illustrates the INTER prediction as proposed in the G-PCC reference software and adopted tools document; Figure 10 is an illustration of the INTER predictive encoding in embodiments of the invention; Figure 11 illustrates the way the INTER predictor index is encoded in the prior art; Figure 12 illustrates an improvement for the coding of the INTER predictor index according to embodiments of the invention; Figure 13 illustrates some details of the improvements for the coding of the INTER predictor index according to embodiments of the invention; Figure 14 illustrates some details of the improvements for the coding of the INTER predictor index according to embodiments of the invention; Figure 15 illustrates another embodiment of the invention; Figure 16 illustrates some details related to an embodiment of the invention; Figure 17 illustrates the decoding step related to prior art in the adopted tool reference software; Figure 18 illustrates the decoding step in an embodiment of the invention; and Figure 19 illustrates a block diagram of a device adapted to incorporate embodiments of the invention. DETAILED DESCRIPTION OF THE INVENTION This disclosure is related to a method for coding the INTER predictor, called refNodeFlag, for the Predictive Geometry. In this disclosure, in the preferred embodiment, it is proposed: • To use two contexts instead of one for the arithmetic encoder in charge of encoding the INTER predictors refNodeFlag. • To set the choice between the two contexts from the presence of a point in the reference frame with same scaled azimuth as the scaled azimuth of previous encoded point in the current frame. This document is related to predictive geometry and INTER prediction for predictive geometry. A set of 2 INTER predictors has been proposed for predictive geometry. These two INTER prediction methods have been added in the list of adopted tools for a future version G-PCC. The first INTER predictor references the function called ‘getClosestPred’ and corresponds to a refNodeFlag value set to false. The second INTER predictor references the function called ‘getNextClosestPred’ and corresponds to a refNodeFlag value set to true. The INTER predictors index, refNodeFlag, is encoded with an arithmetic encoder with one context. In this disclosure, it is proposed to add a second context for the compression of the INTER predictor index. This modification of the source code is very low and has the following characteristics: • Compression ratios are improved. • Coding and decoding times are almost unchanged. The reference software of the prior art for estimating and encoding the INTER prediction index refNodeFlag is based on the following elements: • The usage of two INTER prediction functions, Inter Pred and Additional Inter Pred as illustrated by Figure 9 and 10. • One arithmetic context for the encoding and decoding of refNodeFlag: o AdaptiveBitModel _ctxRefNodeFlag; • Bits estimation of the bit cost of refNodeFlag based on the unique context: o bits += estimate(refNodeFlag, _ctxRefNodeFlag); • Encoding of the refNodeFlag with the unique context: o _aec->encode(refNodeFlag, _ctxRefNodeFlag); In this disclosure, the following modifications are proposed: • 2 arithmetic contexts for the encoder and decoder of refNodeFlag: o AdaptiveBitModel _ctxRefNodeFlag[2]; • A modification of the function in charge of the bits estimation for refNodeFlag: o bits += estimate(interFlag, _ctxlnterFlag[isPreviousPredictorExist]); o isPreviousPredictorExist: is a two states variable defined below. • A modification of the function in charge of the encoding and decoding of refNodeFlag: o _aec->encode(refNodeFlag, _ctxRefNodeFlag[isPreviousPredictorExist]); o _aec->encode(refNodeFlag, _ctxRefNodeFlag); o _aed->decode(refNodeFlag, _ctxRefNodeFlag[isPreviousPredictorExist]); o isPreviousPredictorExist: is a 2 state variable defined later. The state variable “isPreviousPredictorExist” is set to 1 if a point in the reference frame with same scaled azimuth as the scaled azimuth of previous encoded / decoded point in the current frame exists. The state variable “isPreviousPredictorExist” is set to 0 if no point in the reference frame with same scaled azimuth as the scaled azimuth of previous encoded / decoded point in the current frame exists. This is illustrated by the Figure 9. As conclusion, we can say that this disclosure is related to a method for coding the INTER predictor, called refNodeFlag, for the Predictive Geometry. Compression gains are obtained by using the proposed method. Figure 1 illustrates an example of LIDAR system where the 4 arrows 1000 are representative of 4 laser heads. The 4 heads rotate around the z axis at a given rotation speed. Each head emits laser beam regularly and calculate a 3D position (x,y,z coordinates in a Cartesian system) according to measured feedbacks relative to the laser beam emission. In addition, to these 3D positions, attributes may be present for each point. At each 3D position, a set of attributes can be associated. For example, the attributes can be a color, a normal, reflectance, etc. This kind of LIDAR is called rotating LIDAR, but the invention is not limited to this particular example of LIDAR and may be adapted to any kind of 3D dynamic point clouds. In the context of MPEG standardization, a compression standard called G-PCC V1 has emerged and is under finalization. Figure 2 illustrates different compression tools proposed in this standard: In step 2000, the initial point cloud is considered. It corresponds to a frame captured at time ‘t’. This initial point cloud may be optionally pre-processed in step 2001 to quantize the 3D positions of the 3D points. This first step is often called ‘voxelization’. The voxelization moves slightly the 3D positions of the initial point cloud so that they move to the center of the nearest voxel in a bounding box. - A step 2002, called geometry encoding, consists in encoding the 3D positions of the point cloud. G-PCC proposes three alternatives, 2003, 2004 and 2009, for the encoding of the geometry: o The first alternative 2003 is based on the construction of an octree and the encoding of the generated octree structure. o The second alternative 2004 is based on the construction of a tree where each 3D point position is encoded in reference to previously encoded point positions in the tree. This second alternative is called geometry prediction. o The third alternative 2009 is based on the construction of a surface of triangles from an initial partial octree coding. Some parameters are added in the initial octree coding enabling the construction of the surface. Intersections between the surface and voxels enables to generate a point cloud approximating the initial point cloud to encode - Once the geometry is encoded, either from octree-based technics, geometry prediction technics or trisoup, the geometry is decoded to obtain a decoded 3D position of the points. The decoded positions, which can be different from the initial positions if the geometry is encoded with the lossy mode, can be re-colored in step 2005. It means that attributes of the encoded-decoded points can be changed according to the differences between the initial 3D position of a point and its decoded position. For example, the color or normal of each point can be recalculated taking into account the modification between the initial position and decoded position of a point. If the geometry is encoded in lossless mode, there is no need in applying this step 2005. The result of this recoloring is a new set of attributes related to the encoded-decoded 3D points. - After recoloring, the attributes are encoded in a step 2006. G-PCC proposes 2 modes for encoding the attributes called RAHT, step 2007, and LoD / Lifting, step 2008. Usually, the dense point clouds use the octree-based technics (2003) for geometry encoding step 2002, thus for compressing the positions of the point cloud called the geometry. The sparse point clouds use typically the geometry prediction technics of step 2004. Indeed, geometry prediction is more efficient when dealing with sparse content. LIDAR generated point clouds are often sparse. Trisoup is dedicated to low bit-rate and is used for dense point clouds. It consists in encoding the positions of point cloud by transmission of triangular-based surface (parameters of this surface being the compression parameters) and in intersecting the surface with voxels. The intersections will be considered as the reconstructed points. This invention is related to step 2004 of geometry prediction. The contemplated standard G-PCC V1, as illustrated in Figure 2, suffers from drawbacks: One major drawback is the independent temporal compression of the frames of the temporal point cloud. Temporal independency means that the current frame, whose capture starts at time ‘t’, is compressed and encoded independently of the previous frame, whose capture starts at time ‘t-1’. The redundant information between the two frames cannot be exploited. That is the reason why the new technologies under development in the G-PCC context described in the TUC document, has proposed INTER tools for taking advantage of the temporal redundancy and improving the compression ratio. The INTER tools proposed by the TUC document exist both for geometry encoding (2002) and attributes encoding (2006). Recently, this INTER tools have been added in a list of adopted tools for the future version of G-PCC. The predictive encoding of a 3D point consists in determining a predictor of the 3D point to encode. The predictor is the decoded version of a previously encoded point. The predictor is in the current frame, meaning the current 3D point cloud, for INTRA encoding. The predictor is in a previously encoded frame, meaning a previously encoded 3D point cloud, named the reference frame, in case of INTER encoding. The predictor is determined by evaluating a set of predictors and selecting one of the predictors based on this evaluation as the predictor to be used at encoding and decoding. The determination being made by evaluating a cost function for each predictor corresponding to the rate-distortion of the encoding based on the predictor. In the adopted tools reference software, for a given point to encode, two INTER prediction functions are possible. The index (called ‘refNodeFlag’) of the chosen INTER prediction function (if INTER mode is better than INTRA mode), is encoded by an arithmetic encoder based on a simple arithmetic context. If the first INTER prediction function is chosen, ‘refNodeFlag’ is encoded with the Boolean ‘false’ with the arithmetic encoder. If the second INTER prediction function is chosen, ‘refNodeFlag’ is encoded with the Boolean ‘true’ with the arithmetic encoder. According to embodiments of the invention, it is proposed to improve the coding of the ‘refNodeFlag’ value. This will be explained in detail below. The advantage of the new way of encoding the index ‘refNodeFlag’ is an improvement of the compression ratio. Figure 3 illustrates an example of the main steps of a general encoding method for a point cloud and the construction of the associated bitstream. The three main elements are the point cloud / frame 3000 to be encoded, the compression algorithm 3003 which will generate the bitstream 3006. In a step 3001, the point cloud is divided in slices, which are subsets of the 3D points of the point cloud. Each slice is compressed in step 3005 for the geometry and in step 3007 for the 3D points attributes. The compression step 3003, comprises a first step 3004 of initialization, a step of generation of the geometry bitstream 3005, followed by the generation of the attributes bitstream 3007. The generated bitstream 3006 comprises a metadata part 3008, 3009 and 3010 followed by data units 3011-3014. The metadata part comprises a sequence parameter set, SPS, 3008 comprising parameters defined for the whole sequence, a geometry parameter set, GPS, 3009, comprising parameters of the geometry and an attribute parameter set, APS, 3010 comprising parameters of attributes. The data units are of two sorts, geometry data unit, GDU and attributes data unit, ADU. Each data unit is constituted of a header, GDU header 3011 and ADU header 3013, and a payload, GDU data 3012 and ADU data 3014, comprising the raw encoded geometry data and attributes data respectively. The metadata are typically generated during the initialization step 3004, the GDUs during the geometry bitstream generation step 3005, the ADUs during the attributes bitstream generation step 3007. Figure 4 illustrates an example of the main steps of a decoding method for decoding a bitstream generated by the encoder as specified in relation with Figure 3. In a first step 4009, the bitstream 4000 is obtained (e.g. read). The metadata comprising the SPS 4001, the GPS 4002 and the APS 4003, are obtained in a step 4010. These metadata are, for example, read and decompressed and then used in a step 4011 for initializing the Raw data decompression step 4012. The raw data decompression step 4012 is then able to decode the GDlls, 4004, 4005 and the ADlls, 4006, 4007 based on the metadata. Figure 5 illustrates examples of 3D points obtained by a rotating lidar. On the left, the lidar 5000 is rotating clock wise on the ‘z’ axis, based on the reflection of the beam on obstacles, here two cars 5001 and 5002 in front of a wall 5003. The projection on the plane (‘x’, ‘y’) is illustrated. For each point, its coordinates (x, y, z) are obtained. The right part illustrates the same points expressed in spherical coordinates where each point is defined by a laser index corresponding to an elevation angle 0 defined by each laser head geometry as illustrated by Figure 1, the azimuthal angle (rotation angle) q>, and the radius r giving the distance between the lidar and the point. Accordingly, each 3D point may be represented by its spherical coordinates (r, q>, 0). The geometry prediction method uses mainly the spherical coordinates. According to G-PCC-V1, the encoding of geometry is now described. As explained in relation to Figure 2, G-PCC proposes three methods for encoding the geometry of the 3D points. The predictive approach starts by defining a prediction structure on the point cloud. Such structure could be described by a prediction tree where each point in the point cloud is associated with a vertex of the tree. Each vertex could be predicted only from its ancestors in the tree. Various prediction strategies are possible. The standard proposes four different strategies for determining a predictor: no prediction, meaning the current point is encoded with a predefined predictor, delta prediction where the predictor is the direct ancestor of the current point in the tree (i.e. pO), Iinear2 prediction where the predictor is a linear combination of the two ancestors (i.e. 2p0-p1), and Iinear3 prediction where the predictor is a linear combination of the three ancestors (i.e., 2p0+p1-p2), where pO, p1, p2 are the positions of the parent, grandparent, and grand-grandparent of the current vertex. The tree structure is encoded by traversing the tree in a depth order and encoding for each vertex the number of its children. The positions of the vertices are encoded by storing the chosen prediction mode and the obtained prediction residuals. Arithmetic coding is used to further compress the generated values. Building the optimal prediction tree is an NP-hard problem. Prior to generating each predictive tree, the input points corresponding to the tree are sorted according to a sorting method. This helps to guide the tree construction process to build a more efficient tree. The sorting methods available are none, morton order, azimuth angle order, and radial distance order. Especially, the azimuthal sorting by rounded azimuth, radius, and elevation generates stable ordering shape and improves the coding efficiency. In G-PCC’s predictive tree, for each node in the prediction tree, one predictor index ‘n’ is encoded in the bitstream. This index points to a selected predictor PRn among a list of possible predictors, called predictors. When angular mode (spherical coordinates) is used the predictors are PRo, PRi, PR2 and PR3 defined as follows: 1) “None”: PRo = (rmin, <po, 0o), where rmin is the minimum radius value (provided in the geometry parameter set), and (po and 0o are equal to 0 if the node has no parent, or are equal to (p and 0 values of the point coded in the parent node. 2) “Delta”: PR1 = pO = (ro, (po, 0o), where ro, (po and 0o are respectively the radius r, the azimuthal angle (p and the laser index 0 values of the parent point pO coded in the parent node. 3) “Linear2”: PR2 = 2*p0-p1, where pO and p1 are the parent and grandparent points / nodes. 4) “Linear3”: PR3 = p0+p1-p2, where pO, p1 and p2 are the parent, the grandparent and great grandparent points / nodes. A prediction residual (rres, (Pres, 0res) is obtained in the encoder by (Res, (Pres, 0res) = (r, (p, 0) — PRn — (0, k * (Pstep, 0), (1) where PRn is one of the predictors PRo, PRi, PR2 or PR3, and k is a number of azimuthal angle steps ((psteP) to be added to the prediction. The prediction index ‘n’ and the number ‘k’ of azimuthal angle steps are encoded in the bitstream for each node, while the value of ((psteP) is encoded in the geometry parameter set by geom_angular_azimuth_speed_minus1. The residual (rres, (Pres, 0res) is also encoded in the bitstream. (Pres can be additionally quantized according to the radius r before being encoded in the bitstream. In both encoder and decoder side, the coordinates (rdec, (Pdec, ©dec) of points are retrieved by doing the reverse process: a reconstructed point (rdec, (Pdec, 0dec) is obtained by: (rdec, (Pdec, 0dec) = PRn + (0, k * (Pstep, 0) + (hes, (Pres.rec, 0res), where (pres,rec is obtained from the quantized (pres after being inversely quantized. This way of generating predictors is not the most efficient one. In the TUC document and in the document containing adopted tools for a future version of G-PCC, new prediction functions have been proposed, which are now described. Instead of using the list of G-PCC prediction functions as previously explained, a list of N predictors is built from a prediction buffer of N pairs of one radius and one azimuthal angle (rn, (pn). The coding of the predictor index is simply performed using a unary coding with one context per bin index. The derivation of a predictor is performed as follows: If the point being predicted is the first point of the tree (i.e., there is no parent node), the predictor PRo is set equal to (rmin, 0, 0), the other predictors PRn>o are set equal to (0, 0, 0). If the point has a parent point, the predictor PRo is set equal to (ro, tpo, 0o), where 0o is the laser index 0 value of the parent point pO coded in the parent node, and where (ro, q>o) is the first pair in the buffer (as will be understood from the buffer management, it is also equal to respectively the radius r, and the azimuthal angle (p of the parent point pO coded in the parent node); the predictors PRn>o are set equal to (rn, (pn + k* (pstep, 0o), where 0o is the laser index 0 value of the parent point pO coded in the parent node, and where (rn, (pn) is the n-th pair in the buffer, and k equals 0 if | cp0 - <pn | < (Pstep, else k equals the integer division ((p0 - (pn) I (psteP. Since it is better to avoid integer division in decoder, (tpo - (pn) I (psteP is approximated using the divApprox function of G-PCC: k = divApprox((po - (pn, (psteP, 0). The buffer used for the predictors’ derivation is managed as follows. Each pair of the buffer is first initialized to (0, 0). After the (de)coding of a point, the buffer is updated as follows: If the absolute value of (de)coded radius residual rres is higher than a threshold Th, it is considered that the laser has probed a new object. Then a new element (ro, q>o) is inserted in front of the buffer, with ro and tpo the reconstructed radius and the reconstructed azimuthal angle of the (de)coded point. The last element of the buffer is discarded. This is performed by letting the buffer element (rn, (pn) be equal to (rn-i, <pn-i) for n=3 to 1. Then, setting the first buffer element values from the decoded point. If the absolute value of (de)coded rres is not higher than the threshold Th, it is considered that the laser has probed an object present in the buffer. Then, the element of the buffer with index predldx, corresponding to the index of the predictor that has been used for the prediction, is moved to the front of the list and is updated with (ro, (po) the reconstructed radius and the reconstructed azimuthal angle of the (de)coded point. This is performed by letting the buffer elements (rn, (pn) be equal to (rn-i, <pn-i) for n=predldx to 1, then, setting the first buffer element values from the decoded point. - Th is equal to ps.predgeom_radius_threshold_for_pred_list and has been fixed in the encoder to 2048 » ps.geom_angular_radius_inv_scale_log2 with »’ being a bit shift. The parameter ‘geom_angular_radius_inv_scale_log2’ is an inverse scaling factor applied to radius in predictive geometry coding. It is used to modify the precision in bits of the radius values. In addition, some INTER prediction tools have been added in the predictive geometry scheme. These tools will be explained later. Figure 6 illustrates an example of tree construction for a rotating lidar. Input 3D points 6007 in spherical coordinates (Rj, q>i, 0j), ‘i’ being the index of the point. For rotating Lidar, there is a direct correspondence between elevation 0j and laserld. This is the reason why we use indifferently these 2 words. Before starting the encoding, the tree is constructed in step 6006. The tree construction consists in sorting the points in an increasing azimuth order for a same given elevation except for a short subset of 3D points. This short subset of points is illustrated with the points 6000 and 6003. As these are the first encoded points for a given (different) elevation, their ancestors have a different elevation and there is no guaranty the azimuth of the ancestor is lower. For example, in the tree 6008, the point 6001 has the point 6000 as parent. The point 6002 has the point 6001 as parent. It means that the azimuth <p1 is higher than <p0. The azimuth q>2 is higher than <p1. The point 6003 has the point 6000 as parent because it is the 3D point at elevation 01 with the lowest azimuth. The first point with an elevation higher than 00 (6003) has as parent the point 6000. The point 6004 has the point 6003 as parent. The point 6005 has the point 6004 as parent. In the TUC document, the generated tree is used for: Successively selecting the 3D points to encode (a point being a node of the tree) Use the parent as predictor of the elevation. The radius and azimuth are predicted from the new INTRA and INTER prediction functions. In a preferred embodiment, we suppose that this global prediction mode is kept in this document even if the INTRA or INTER prediction functions could be also used for elevation prediction, or just for radius or just azimuth prediction. Figure 7 illustrates the main step of an example of the compression algorithm used for encoding the 3D geometry of the input points corresponding to step 2004 in Figure 2. This version of the algorithm is used, for example, for a 3D point cloud coming from LIDAR or any 3D point cloud with a spherical representation. As explained previously, a tree is constructed. According to this tree, a point is selected in step 7000. This point is in its spherical representation (radius, azimuth, elevation or Laserld). In step 7001, a prediction of this point is done. The prediction can be done from INTRA prediction functions (figure 5), or INTER prediction functions (figures 8, 9 and 10). The predictors (calculated from used prediction functions) are previous reconstructed 3D points, reconstructed meaning that these points have been encoded and decoded, coming from a reference frame for INTER prediction or from the current frame for INTRA prediction. The best predictor is chosen among all the predictors. A flag is set to notify if it is an INTRA or INTER predictor. The index of this predictor (INTER or INTRA) is encoded and added in the bitstream in step 7006. The predictors are used for the prediction of radius and azimuth. The elevation being predicted from the parent in the tree. Once the prediction is done, the residual is calculated in the spherical representation in step 7002, quantized and encoded in a binary representation in step 7005 for being added in the bitstream in step 7006. The decoded residual in the spherical domain is calculated in step 7003 and added to the prediction in the step 7007. The reconstructed point is transformed in the Cartesian domain / representation in step 7008. In step 7013, the input point is transformed in Cartesian coordinates. The residual in the Cartesian domain is obtained in step 7004. This residual is quantized and encoded in a binary stream in step 7006. In step 7009, the residual is decoded in the Cartesian domain before being added in step 7010 to the point in step 7008. The result is the reconstructed point in the Cartesian domain which is transformed in the spherical domain in step 7011 before being stored in step 7012. The storage in used in step 7012 contains encoded / decoded points that can be used as INTRA or INTER prediction. For INTRA prediction, the encoded / decoded points are used as described above for constructing the list of predictors. For INTER prediction, the stored point will be used for the encoding of the next frame: they will be used as reference frame for the next frame as illustrated by Figure 10. Figure 8 illustrates a rate distortion algorithm. This figure is done for the coding of all the points of the point cloud. In 8000, the INTRA predictors are tested. In 8001, for each tested INTRA predictor, the estimate bits function estimates the coding cost in bits of the tested INTRA predictor. In 8002 and 8003, the same operations are done for the INTER predictors. In 8004, the best predictor (among the INTRA and INTER predictors) is selected and encoded in 8005 for generating the bitstream. The associated residual information is also encoded in 8005. In particular, the bit estimation 8003 requires to estimate the coding cost of the chosen INTER prediction function, the index of this chosen INTER function being called refNodeFlag. In particular, the bit encoding 8005 requires to encode the chosen INTER prediction function, the index of this chosen INTER function being called refNodeFlag. According to an aspect of the invention, it is proposed to change these two operations related to the INTER mode, for the refNodeFlag value, by adding a new context. It means that the proposed solution yields to modification of the Estimate Bits module 8003 for the part related to refNodeFlag and to the modification of the encoding module 8005 for the coding of the refNodeFlag. This will be illustrated later. Some associated modifications will be done for the decoding part for being compatible with the modifications made by the encoder. Figure 9 illustrates the INTER prediction as proposed in a first and second version of the TUC document and in the adopted tools reference software. When inter-coding is applied, the radius, azimuth and laserlD of the current point are predicted based on a point that is near the azimuth position of a previously encoded / decoded point in the reference frame. The reference frame is an already encoded / decoded frame. This frame may have been compensated in motion. This improved method is illustrated in Figure 9. The method consists of the following steps: • For a given 3D point in the frame at time ‘t’, called Curr Point 9000 in the figure, choose the previous encoded / decoded point, Prev Dec Point 9001 in the same frame. • Potentially quantize the azimuth of the previous encoded / decoded point (quantized azimuth can be called ‘scaled azimuth’ or ‘phi quantized’). The potentially quantized value is called phiQ. • Determine a reference point 9004 corresponding to the previous decoded point 9001 in the reference frame. This reference point 9004 has typically the same azimuth phiQ than the previous decoded point 9001. Depending on the resolution of each frame, this reference point 9004 with the same azimuth phiQ may exist or not in the reference frame. • Choose position in the reference frame a first prediction point (called Inter Pred Point 9002 in the figure). The prediction point is the prediction point of the reference frame with same elevation and first quantized azimuth higher than phiQ. This is the point following the reference point 9004 in the reference frame when this reference point 9004 exists. • Choose position in the reference frame a second prediction point (called Additional Inter Pred Point 9003 in the figure). The prediction point is the prediction point of the reference frame with same elevation and second quantized azimuth higher than phiQ. When the INTER predictor is chosen, a flag is encoded, called “inter_flag”, for signalling to the decoder to use INTER predictor. If the first INTER prediction point is chosen, a flag called “refNodeFlag’ is set to false. If the second INTER prediction point is chosen, the flag called ‘refNodeFlag’ is set to true. This flag value is next encoded with an arithmetic encoder. In the state of the art, the arithmetic encoder described later in relation to Figure 11 is based on one single arithmetic context. The flag ‘refNodeFlag’ indicates which predictor is selected between the first and the second INTER predictor. According to an embodiment of the invention, it is proposed to improve the encoding of the flag called ‘refNodeFlag’ by using an arithmetic context different from the one used in the state of the art. Figure 10 is another illustration of the INTER predictive encoding described above in relation to Figure 9. The figure illustrates the current frame 10000 and the reference frame 10001. The reference frame 10001 could be any previously encoded frame. The reference frame could be translated and rotated according to 3D parameters representing the 3D motion between the reference frame and the current frame, these parameters being transmitted in the bitstream for the decoder. For the current and reference frame, several points with same elevation (00) are illustrated. The notion of elevation and laserld are equivalent because the elevation can be calculated from the laserld. In this document, we use the term elevation, but Laser ID can be used instead for rotating LIDAR. In real frames, points at different elevations exist because the LIDAR has several heads as illustrated in Figure 1 but the representation considers only one elevation (or laserld) for simplicity. As indicated in Figure 6, a tree can be constructed (and the points to encode successively selected according to this tree) according to increased azimuth for a given elevation. The method is applied similarly for each elevation. The point to encode 10002 is illustrated for the given elevation. The point is in the spherical representation wherein q>3 is the value of the azimuth, R3, the value the radius and 00 is the elevation. The previous encoded / decoded point 10003 of the same frame is selected according to the constructed tree. The azimuth value of this point may be quantized. In this figure the associated quantization value is called phiQ.lf there is no quantization, the quantized azimuth equals the azimuth. The reference point 10004 collocated in the reference frame is selected and the two INTER predictors 10005 and 10006 are evaluated. The reference point is a point of the reference frame with same quantized azimuth. In this example, let’s consider that the additional predictor 10006 is the best one. This predictor could be used as predictor of the current point 10002 for the prediction of the radius, azimuth and elevation. If the reference point 10004 doesn’t exist because no feedback from the laser beam has been received, the two INTER predictors are the first two points of the reference frame with same elevation and quantized azimuth higher than phiQ 10005 and 10006. If one of the two INTER predictors 10005 and 10006 is better, according to a ratedistortion estimate, than the INTRA predictor, then the flag signalling the choice between INTRA or INTER mode is set for INTER. And the flag indicating the selected INTER predictor, refNodeFlag, is encoded as a Boolean value. Figure 11 illustrates the state of the art for the coding of the refNodeFlag. Indeed, if the INTER mode is better than the INTRA mode, the choice of the INTER prediction function (refNodeFlag: true or false) has to be encoded for being used at the decoder. In addition, it is explained how the estimated cost of the refNodeFlag coding is done. The estimation of the cost of the encoding of refNodeFlag is necessary during the bit rate optimization as described in the Figure 8. In 11000, the refNodeFlag is taken as input of the arithmetic encoder 11001. In 11002, the arithmetic encoder called ‘_aec’ in the exemplary embodiment makes use of the encode method 11002 by taking 2 arguments: refNodeFlag and _ctxrefNodeFlag. _ctxrefNodeFlag is the context associated with the arithmetic encoder and the probability data used for the arithmetic encoder are stored in an object called AdaptiveBitModel. The context is a kind of probability table (probability data) gathering probabilities associated with already encoded / decoded refNodeFlag values. This kind of arithmetic encoder is well known in the state of the art. A description is provided at the end of the document for this kind of encoder. After encoding 11002, the coded data associated with the value refNodeFlag are added in the bitstream. During the rate distortion optimization (as described in the Figure 8), the different tested refNodeFlag (11004) are used as input of the estimate bits block 11005 and their coding cost is evaluated by the function estimate 11006 with 2 arguments refNodeFlag and _ctxrefNodeFlag. The context ‘_ctxrefNodeFlag’ is the same as the one used by the arithmetic encoder 11001. The estimated cost is added to the variable bits which is the variable for estimating the costs of all the information (e.g. the different flags, residual, ....). The efficiency of an arithmetic encoder depends on the relevance of the probability of the different values to encode as provided by the arithmetic context. The inventors have found out that the probability of the refNodeFlag values, here zero or one, are not the same depending on a criterion which is the existence of the reference point 10004 having the same azimuth phiQ than the previously encoded / decoded point 10003. Accordingly, it is proposed to use two different contexts for encoding refNodeFlag, the choice of the context depending on the existence of this reference point. In some embodiments, more than two different contexts may be used as illustrated by Figure 16. The choice of the contextis based on classifying the points to be encoded into different classes of point based on some criteria. In each class, the statistic distribution of the values of refNodeFlag is more unbalanced (or biased) than the one used if only one class (one context) is used. For better explaining this notion of unbalance (bias), we can take an example. Let us suppose a binary source with equal probability of 0 and 1. As the probability of 0 and 1 is 0.5, the arithmetic encoding efficiency is very poor. The probability of 0 or 1 may change depending on whether a given condition is fulfilled or not. Because the conditional probability distribution is more unbalanced, the Shannon entropy decreases and the arithmetic encoding using the given condition becomes more efficient. The exemplary embodiment comprises only two possible predictors and a flag to indicate which of these 2 predictors is selected at encoding. The proposed solution may be extended to a set of predictors comprising different number of predictors. The Boolean refNodeFlag is the index of the selected predictor at encoding. The arithmetic encoding of this index may use the classification of the 3D points to be encoded into a number of classes for improved performances. One arithmetic context (probability model of the bins according to the context) is then maintained for each class and used for the encoding of the indexes of 3D points belonging to the class. Figure 12 illustrates an embodiment of the invention for the coding of the refNodeFlag. This embodiment is in relation to the coding of the index of the INTER prediction function refNodeFlag which can take two different values true or false. In particular, it is related to the arithmetic encoding of the refNodeFlag. The main difference with the Figure 11 is the context used by the arithmetic encoder. In this embodiment, the context is based on 2 objects AdaptiveBitModel 12003. With this modification, the performances of the encoding of the INTER prediction function flag refNodeFlag are improved. In 12000, the refNodeFlag is taken as input of the arithmetic encoder 12001. In 12002, the arithmetic encoder called ‘_aec’ makes use of the encode method 12002 by taking 2 arguments: refNodeFlag and _ctxrefNodeFlag[isPreviousPredictorExist]. The difference with the state of the art comes from the parameter ‘isPreviousPredictorExist’. This value is a Boolean value. If this value is false, it means that the first class is associated with the current point, meaning that the first context is used for the encoding. If the value is true, it means that the second class is associated with the current point, meaning that the second context is used for the encoding. As isPreviousPredictorExist may take two values, there are two possible classes in relation with the two contexts. The parameter or variable ‘isPreviousPredictorExist’ is used as criterion for determining the class associated with the current point. The way this variable is calculated is explained later in relation to figure 13. If the value of this variable is false, the arithmetic encoder will use the context _ctxrefNodeFlag[0]. If the value is true, the arithmetic encoder will use the context _ctxrefNodeFlag[1], After encoding 12002, the coded data associated with the value refNodeFlag are added in the bitstream 12007. During the rate distortion optimization, the different tested refNodeFlag 12004 are used as input of the estimate bits block 12005 and their coding cost is evaluated by the function estimate 12006 with 2 arguments refNodeFlag and _ctxrefNodeFlag[isPreviousPredictorExist]. The context ‘_ctxrefNodeFlag’ is the same as the one used by the arithmetic encoder 12001. The estimated cost is added to the variable bits which is the variable for estimating the costs of all the information (e.g. flags, residual, ....) associated with a given prediction. The variable isPreviousPredictorExist is calculated in the same way both for the bit estimation and the encoding. Figure 13 illustrates the way the variable ‘isPreviousPredictorExist’ is calculated. This is the criterion used to calculate the class associated with the current point. This variable is calculated one time before the rate-distortion optimization and the refNodeFlag encoding and used as explained in the previous figure during the bit-rate optimization and the encoding. This variable is calculated for each new point to encode. 13000 is the current frame containing points to encode. The points are sorted according to a prediction tree as illustrated before. 13001 is a reference frame previously encoded / decoded containing points sorted according to the prediction tree. The reference frame is a frame potentially compensated in motion. The point to encode 13002 is represented in spherical representation wherein q>2 is the value of the azimuth, R2, the value the radius and 00 is the elevation. The previous encoded / decoded point of the same frame 13003 is selected. At the decoder, this point is the previously decoded point. The azimuth <p1 of the previous point is read. This azimuth is quantized in 13004 conducting to the value ‘phiQ_prev’. in 13005, it is searched in the reference frame if a corresponding point exists. A corresponding point is a point with same elevation and whose quantized azimuth value is equal to ‘phiQ_prev’. For example, if the point 13006 in the reference frame has a quantized azimuth quantization(q)T) equal to ‘phiQ_prev’ then the response 13007 to the searching 13005 will be positive and the variable ‘isPreviousPredictorExist’ is set to 1. It means that the current point is associated with the second class. In case of no corresponding point is found, the variable ‘isPreviousPredictorExist’ is set to 0. It means that the current point is associated with the first class. This second scenario is illustrated in the Figure 14. Figure 14 illustrates the way the variable ‘isPreviousPredictorExist’ is calculated. This variable is calculated one time before the rate-distortion optimization and the refNodeFlag encoding and used during the bit-rate optimization and the encoding. This variable is calculated for each new point to encode. 14000 is the current frame containing points to encode. The points are sorted according prediction tree as illustrated before. 14001 is a reference frame containing points sorted according to tree. The reference frame is a frame already encoded / decoded potentially compensated in motion. The point to encode 14002 is represented in spherical representation wherein <p2 is the value of the azimuth, R2, the value the radius and 00 is the elevation. The previous encoded / decoded point of the same frame 14003 is selected (at the decoder, this point is the previous decoded point). The azimuth <p1 of the previous point is read. This azimuth is quantized in 14004 conducting to the value ‘phiQ_prev’. In 14005, it is searched in the reference frame if a corresponding point exists. A corresponding point is a point with same elevation and whose quantized azimuth value is equal to ‘phiQ_prev’. For example, if the point 14006 in the reference frame has a quantized azimuth quantization(q)T) equal to ‘phiQ_prev’ then the response 14007 to the searching 14005 will be positive and the variable ‘isPreviousPredictorExist’ is set to 1. In case of no corresponding point is found, the variable ‘isPreviousPredictorExist’ is set to 0. This is the case in this figure wherein we can see that none of the quantization values quantization(q)O’) and quantization^’) corresponding to the two points 14006 has a value close to the value ‘phiQ_prev’. Figure 15 illustrates a second embodiment of the invention for the coding of the refNodeFlag. This invention is in relation to the encoding of the flag refNodeFlag. In particular, it is related to the arithmetic encoding of the flag refNodeFlag. The main difference with Figure 12 is the context used by the arithmetic encoder. Now the context is based on 3 objects AdaptiveBitModel 15003. This means that three different contexts are maintained for the encoding of the flag. Actually, the choice between the three different contexts is based on the existence in the reference frame of reference points collocated with the two previous decoded points in the current frame. With this modification, the performances of the INTER mode are improved. In 15000, the flag refNodeFlag is taken as input of the arithmetic encoder 15001. In 15002, the arithmetic encoder, called ‘_aec’, makes use of the encode method 15002 by taking 2 arguments: refNodeFlag and_ctxrefNodeFlag[isPreviousPredictorExist], The difference with the state of the art comes from the ‘isPreviousPredictorExist’. This value is an integer value, the way it is calculated is explained later. If the value is 0, the arithmetic encoder will use the context_ctxrefNodeFlag[0]. If the value is 1, the arithmetic encoder will use the context _ctxrefNodeFlag[1], If the value is 2, the arithmetic encoder will use the context _ctxrefNodeFlag[2], It means that the current point to encode can be associated with 3 classes. The association with one of these classes conduct to the choice of the related context. After encoding 15002, the coded data associated with the value refNodeFlag are added in the bitstream 15007. During the rate distortion optimization, the different tested refNodeFlag values 15004 are used as input of the estimate bits block 15005 and their coding cost is evaluated by the function estimate 15006 with 2 arguments the refNodeFlag value to encode and _ctxrefNodeFlag[isPreviousPredictorExist], The context ‘_ctxrefNodeFlag’ is the same as the one used by the arithmetic encoder 15001. The estimated cost is added to the variable bits which is the variable for estimating the costs of all the information (e.g. flags, residual, ....). Figure 16 illustrates a second embodiment for the calculation of the variable ‘isPreviousPredictorExist’. This variable is calculated one time before the rate-distortion optimization and the refNodeFlag value encoding and used as explained in the previous figure during the bit-rate optimization and the encoding. This variable is calculated for each new point to encode. 16000 is the current frame containing points to encode. The points are sorted according to a prediction tree as illustrated before. 16001 is a reference frame containing points sorted according to the prediction tree. The reference frame is a frame already encoded / decoded potentially compensated in motion. The point to encode 16002 is represented in spherical representation wherein q>2 is the value of the azimuth, R2, the value the radius and 00 is the elevation. The previous encoded / decoded point of the same frame 16003 is selected (at the decoder, this point is the previous decoded point). The azimuth <p1 of the previous point is read. This azimuth is quantized in 16004 conducting to the value ‘phiQ_prev’. The previous previous encoded / decoded point of the same frame 16008 is selected. At the decoder, this point is the previous previous decoded point. The azimuth <p0 of the previous previous point is read. This azimuth is quantized in 16004 conducting to the value ‘phiQ_prev_prev’. In 16005, it is searched in the reference frame if 2 corresponding points exist. A corresponding point is a point with same elevation and whose quantized azimuth value is equal to ‘phiQ_prev’ or to ‘phiQ_prev_prev’. For example: If the point 16006 in the reference frame has a quantized azimuth quantization(q)T) equal to ‘phiQ_prev’ and the point 16007 in the reference frame has a quantized azimuth quantization(<p0’) equal to ‘phiQ_prev_prev’ then the response 16009 to the searching 16005 will be positive and the variable ‘isPreviousPredictorExist’ is set to 2. It is the case wherein 2 corresponding points have been found. In case of no corresponding point is found (both for ‘phiQ_prev’ and ‘phiQ_prev_prev’), the variable ‘isPreviousPredictorExist’ is set to 0. In case of one corresponding point is found (either for ‘phiQ_prev’ or ‘phiQ_prev_prev’), the variable ‘isPreviousPredictorExist’ is set to 1. Figure 17 illustrates the state of the art of a method for decoding the refNodeFlag. In 17000, the bitstream is read by the point cloud decoder. The arithmetic decoder called ‘_aed’ calls the decoder method with 1 parameters: the context of the arithmetic decoder ‘_ctxRefNodeFlag’ 17002. The type of this context is called AdaptiveBitModel and is the same as the one used by the encoder as previously described in relation with Figure 11. This context is built in the same way at the encoder and decoder from the statistics of the already encoded / decoded refNodeFlag values at the encoder, from the statistics of the already decoded refNodeFlag values at the decoder. Once the flag value is decoded, ‘refNodeFlag’ can be used by the decoder for extracting the right INTER predictor. Figure 18 illustrates the embodiment of the invention for the decoding of the refNodeFlag. This decoding has to be understood in reference to the encoding stage of refNodeFlag as explained in relation with Figure 12. In 18000, the bitstream is read by the point cloud decoder (for the decoding of refNodeFlag). The arithmetic decoder called ‘_aed’ calls the decoder method with 1 parameters: the context of the arithmetic decoder ‘_ctxRefNodeFlag[isPreviousPredictorExist]’ (18002). The type of this context is called AdaptiveBitModel and is the same as the one used by the encoder (Figure 12). This context is built in the same way at the encoder and decoder from the statistics of the already encoded / decoded refNodeFlag at the encoder, from the statistics of the already decoded refNodeFlag at the decoder. The value ‘isPreviousPredictorExist’ is calculated in a symmetric way both in the encoder and in the decoder as illustrated in relation with Figure 13. At the encoder, this calculation is based on the previously encoded / decoded points. At the decoder, this calculation is based on the previously decoded points. Once the flag is decoded, ‘refNodeFlag’ can be given to the decoder for extracting the right INTER predictor. Figure 19 illustrates a block diagram of a device adapted to incorporate embodiments of the invention. Preferably, the device comprises a central processing unit (CPU) 19001 capable of executing instructions from program ROM 19003 on powering up of the receiving device, and instructions relating to a software application from main memory 1002 after the powering up. The main memory 19002 is for example of Random Access Memory (RAM) type which functions as a working area of CPU 19001, and the memory capacity thereof can be expanded by an optional RAM connected to an expansion port (not illustrated). Instructions relating to the software application may be loaded to the main memory 19002 from the hard-disc (HD) 19006 or the program ROM 19003 for example. Such software application, when executed by the CPU 19001, causes the steps of the flowcharts shown in the previous figures. Reference numeral 19004 is a network interface that allows the connection of the device to the communication network. The software application when executed by the CPU is adapted to receive data streams through the network interface from other devices. Reference numeral 19005 represents a user interface to display information to, and / or receive inputs from a user. In this section, we are going to describe roughly how arithmetic encoding works. In particular we are going to describe context adaptive binary arithmetic encoder used in embodiments of the invention. Binary arithmetic encoding consists in reading a sequence of bins (sequences of 0 and 1) and in generating a floating interval (for each bin, a new floating interval is calculated based on the previous calculated interval) between zero and one. The interval is constructed based on probability of bins (probability of zero or one). When the entire sequence of bins is processed, this interval can be identified by a floating value. This floating value is represented in a binary form (the bits of the bitstream) and can be used by the arithmetic decoder for decoding the bins. It is supposed that the probabilities of bins are known by the decoder. With this means, it is possible to have code words (bitstream) close to the Shannon entropy. For example, a 20 binary bins can be encoded (for some adequate probabilities) in 16 bits or less. For practical applications of arithmetic encoder, there is no need to wait the entire sequence of bins to be encoded before generating bits for the bitstream. Context Adaptive Binary arithmetic encoding is an improvement over binary arithmetic encoding. It uses additional contexts. Contexts enable to specify ‘sharper’ probability tables. In other words, context allows probability modelling for specific conditions. These specific conditions are equivalent to conditional probabilities. If the selected condition enables to increase the probability of a binary value (for example if the probability of 1 increases when a condition ‘A’ is true), it is interesting to use this condition state. Indeed, as the probability of the binary value increases, the Shannon entropy decreases and the arithmetic encoder compresses better (because arithmetic encoder is able to generate bits / bitstream close to the Shannon entropy). The probability model used by the encoder and decoder (for a given bin) can be known in advance or can be calculated on the flow. For example, the probability model can be constructed from previously decoded elements. In such a way, the probability model will be adaptive and will match as closely as possible the evolution of the probability distribution of the given bins to encode (the more the probability model is close from the real probability, the higher efficient is the arithmetic codec). In an embodiment of the invention, it is proposed to encode with a Context Adaptive Binary Arithmetic encoder the bins related to refNodeFlag. The choice of the contexts is based on the value called ‘isPreviousPredictorExist’ as explained in relation with Figure 15. Any step of the algorithms of the invention may be implemented in software by execution of a set of instructions or program by a programmable computing machine, such as a PC (“Personal Computer”), a DSP (“Digital Signal Processor”) or a microcontroller; or else implemented in hardware by a machine or a dedicated component, such as an FPGA (“Field-Programmable Gate Array”) or an ASIC (“Application-Specific Integrated Circuit”). Although the present invention has been described hereinabove with reference to specific embodiments, the present invention is not limited to the specific embodiments, and modifications will be apparent to a skilled person in the art which lie within the scope of the present invention. Many further modifications and variations will suggest themselves to those versed in the art upon making reference to the foregoing illustrative embodiments, which are given by way of example only and which are not intended to limit the scope of the invention, that being determined solely by the appended claims. In particular the different features from different embodiments may be interchanged, where appropriate. Each of the embodiments of the invention described above can be implemented solely or as a combination of a plurality of the embodiments. Also, features from different embodiments can be combined where necessary or where the combination of elements or features from individual embodiments in a single embodiment is beneficial. In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. The mere fact that different features are recited in mutually different dependent claims does not indicate that a combination of these features cannot be advantageously used.
Claims
1. A method of encoding a 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, in a bitstream, wherein the method comprises for encoding a current 3D point in a current frame of 3D points:- determining a set of predictors in a reference frame that may be used for predictive encoding of the current 3D point;- determining an index of a selected predictor in the set of predictors, the selected predictor being used for encoding the current 3D point;- determining a class in a set of classes to which the current 3D point belongs based on a criterion; and- encoding, using a context based on a probability distribution of previously encoded indexes in the determined class, the determined index.
2. The method of claim 1, wherein the criterion is based on the existence of one or more collocated point in the reference frame corresponding respectively to one or more previously encoded point in the current frame.
3. The method of claim 2, wherein the criterion is based on the existence of a collocated point in the reference frame corresponding to a previously encoded point in the current frame.
4. The method of claim 1, wherein the set of predictors point comprises only two predictors, and wherein the index is a Boolean value indicating which predictor is the selected predictor.
5. A method of decoding a bitstream comprising an encoded 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, wherein the method comprises for decoding a current 3D point in a current frame of 3D points:- determining a set of predictors points in a reference frame that may be used for predictive encoding of the current 3D point;- determining a class in a set of classes to which the current 3D point belongs based on a criterion; and- decoding an index of a selected predictor in the set of predictors, from the bitstream, using a context based on a probability distribution of previously encoded indexes in the determined class.
6. A device for encoding a 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, in a bitstream, wherein the device comprises for encoding a current 3D point in a current frame of 3D points a processor configured for:- determining a set of predictors in a reference frame that may be used for predictive encoding of the current 3D point;- determining an index of a selected predictor in the set of predictors, the selected predictor being used for encoding the current 3D point;- determining a class in a set of classes to which the current 3D point belongs based on a criterion; and- encoding, using a context based on a probability distribution of previously encoded indexes in the determined class, the determined index.
7. A device for decoding a bitstream comprising an encoded 3D dynamic point cloud, comprising a sequence of frames of 3D point clouds, each 3D point cloud comprising a set of 3D points, wherein the device comprises for decoding a current 3D point in a current frame of 3D points a processor configured for:- determining a set of predictors points in a reference frame that may be used for predictive encoding of the current 3D point;- determining a class in a set of classes to which the current 3D point belongs based on a criterion; and- decoding an index of a selected predictor in the set of predictors, from the bitstream, using a context based on a probability distribution of previously encoded indexes in the determined class.
8. A computer program product for a programmable apparatus, the computer program product comprising a sequence of instructions for implementing a method according to any one of claims 1 to 5, when loaded into and executed by the programmable apparatus.
9. A computer-readable storage medium storing instructions of a computer program for implementing a method according to any one of claims 1 to 5.
10. A computer program which upon execution causes the method of any one of5 claims 1 to 5 to be performed.
Citation Information
Patent Citations
Techniques and apparatus for interframe point-cloud attribute coding
US20200304823A1
Predictive geometry coding in g-pcc
US20220108491A1
Point cloud data transmission method, point cloud data transmission device, point cloud data reception method, and point cloud data reception device
WO2022080982A1
Hybrid-tree coding for inter and intra prediction for geometry coding
WO2022147015A1
