Apple Patent | Point cloud compression with multi-layer projection

Patent: Point cloud compression with multi-layer projection

Drawings: Click to check drawins

Publication Number: 20210012536

Publication Date: 20210114

Applicant: Apple

Assignee: Apple Inc.

Abstract

A system comprises an encoder configured to compress attribute information and/or spatial for a point cloud and/or a decoder configured to decompress compressed attribute and/or spatial information for the point cloud. To compress the attribute and/or spatial information, the encoder is configured to convert a point cloud into an image based representation. Also, the decoder is configured to generate a decompressed point cloud based on an image based representation of a point cloud. The encoder is configured project the point cloud on to patch planes to compress the point cloud, and supports multiple layered patch planes. For example, some point clouds may have a depth, and points at different depths may be assigned to different layered patch planes.

Claims

1.-20. (canceled)

21. A non-transitory computer-readable medium storing program instructions that, when executed by one or more processors, cause the one or more processors to: receive encoded point cloud information comprising: a patch image comprising a first set of points of the point cloud corresponding to a patch projected onto a patch plane; and another patch image comprising another set of points of the point cloud, at a different layer of the point cloud, corresponding to the patch projected on the patch plane; decode the encoded point cloud information; and generate a reconstructed point cloud based on the decoded point cloud information.

22. The non-transitory computer-readable medium of claim 21, wherein the points of the other set of points comprise points that are located at respective same pixel locations in the patch plane as points of the first set of points.

23. The non-transitory computer-readable medium of claim 21, wherein the patch image comprising the other set of points of the point cloud at the different layer has been down-scaled relative to the patch image comprising the first set of points of the point cloud.

24. The non-transitory computer-readable medium of claim 23, wherein the program instructions, when executed on the one or more processors, further cause the one or more processors to: up sample the other set of points to determine additional points at the different layer of the point cloud to be used in generating the reconstructed point cloud.

25. The non-transitory computer-readable medium of 24, wherein the up sampling of the other set of points is performed using point information from another layer of the point cloud comprising the set of points included in the patch image.

26. The non-transitory computer-readable medium of 24, wherein the up sampling of the other set of points is performed using motion estimation.

27. The non-transitory computer-readable medium of 21, wherein the patch image and the other patch image are included in different frames, and wherein said decoding the encoded patch information comprises decoding the different image frames in parallel.

28. A device, comprising: a memory storing program instructions; and one or more processors, wherein the program instructions, when executed on the one or more processors, cause the one or more processors to: receive encoded point cloud information comprising: a patch image comprising a first set of points of the point cloud corresponding to a patch projected onto a patch plane; and another patch image comprising another set of points of the point cloud, at a different layer of the point cloud, corresponding to the patch projected on the patch plane; decode the encoded point cloud information; and generate a reconstructed point cloud based on the decoded point cloud information.

29. The device of claim 28, further comprising: a display, wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: render at least a portion of the reconstructed point cloud in the display of the device.

30. The device of claim 28, wherein the points of the other set of points comprise points that are located at respective same pixel locations in the patch plane as points of the first set of points.

31. The device of claim 28, wherein the patch image comprising the other set of points of the point cloud at the different layer has been down-scaled relative to the patch image comprising the first set of points of the point cloud, and wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: up sample the other set of points to determine additional points at the different layer of the point cloud to be used in generating the reconstructed point cloud.

32. The device of claim 31, wherein the up sampling of the other set of points is performed using point information from another layer of the point cloud comprising the set of points included in the patch image.

33. The device of claim 31, wherein the up sampling of the other set of points is performed using motion estimation.

34. The device of claim 28, wherein the patch image and the other patch image are included in different frames, and wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: decode the different image frames in parallel.

35. A non-transitory computer-readable medium storing program instructions that, when executed by one or more processors, cause the one or more processors to: determine, for a point cloud, a plurality of patches each corresponding to portions of the point cloud; for respective ones of the patches, generate a patch image comprising a first set of points of the point cloud corresponding to the patch projected onto a patch plane; generate an additional patch image comprising another set of points of the point cloud, at a different depth of the point cloud, corresponding to the patch projected on the patch plane; pack the generated patch images for the respective ones of the patches into one or more image frames; and encode the one or more image frames.

36. The non-transitory computer-readable medium of claim 35, wherein the points of the other set of points comprise points that are located at respective same pixel locations in the patch plane as points of the first set of points.

37. The non-transitory computer-readable medium of claim 35, wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: assign points of the point cloud within a first depth distance range from the patch plane to the first set of points used in generating the patch image; and assign points of the cloud within a second depth distance range from the patch plane to the set of points used in generating the additional patch image.

38. The non-transitory computer-readable medium of claim 35, wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: determine a representative geometry value or attribute value for the assigned points within the first depth distance; and determine a representative geometry value or attribute value for the assigned points within the additional depth distance, wherein the representative geometry values or attribute values are used in generating the patch image and the additional patch image.

39. The non-transitory computer-readable medium of claim 35, wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: down sample the points within the second depth distance range, such that the generated additional patch image includes a down-sampled representation of the points within the second depth distance range.

40. The system of claim 35, wherein the patch image and the additional patch image are packed into different image frames, and wherein the program instructions, when executed by the one or more processors, further cause the one or more processors to: encode the different image frames using different image resolutions.

Description

[0001] This application is a continuation of U.S. patent application Ser. No. 16/198,634, filed Nov. 21, 2018, which claims benefit of priority to U.S. Provisional Application Ser. No. 62/590,191, entitled “Point Cloud Compression with Multi-Lay Projection,” filed Nov. 22, 2017, and which is incorporated herein by reference in its entirety.

BACKGROUND

Technical Field

[0002] This disclosure relates generally to compression and decompression of point clouds comprising a plurality of points, each having associated spatial information and attribute information.

Description of the Related Art

[0003] Various types of sensors, such as light detection and ranging (LIDAR) systems, 3-D-cameras, 3-D scanners, etc. may capture data indicating positions of points in three dimensional space, for example positions in the X, Y, and Z planes. Also, such systems may further capture attribute information in addition to spatial information for the respective points, such as color information (e.g. RGB values), texture information, intensity attributes, reflectivity attributes, motion related attributes, modality attributes, or various other attributes. In some circumstances, additional attributes may be assigned to the respective points, such as a time-stamp when the point was captured. Points captured by such sensors may make up a “point cloud” comprising a set of points each having associated spatial information and one or more associated attributes. In some circumstances, a point cloud may include thousands of points, hundreds of thousands of points, millions of points, or even more points. Also, in some circumstances, point clouds may be generated, for example in software, as opposed to being captured by one or more sensors. In either case, such point clouds may include large amounts of data and may be costly and time-consuming to store and transmit.

SUMMARY OF EMBODIMENTS

[0004] In some embodiments, a system includes one or more sensors configured to capture points that collectively make up a point cloud, wherein each of the points comprises spatial information identifying a spatial location of the respective point and attribute information defining one or more attributes associated with the respective point.

[0005] The system also includes an encoder configured to compress the attribute and/or spatial information of the points. To compress the attribute and/or spatial information, the encoder is configured to determine, for the point cloud, a plurality of patches each corresponding to portions of the point cloud, wherein each patch comprises points with surface normal vectors that deviate from one another less than a threshold amount. The encoder is further configured to, for each patch, generate another patch image comprising the set of points corresponding to the patch projected onto a patch plane and generate a patch image comprising depth information for the set of points corresponding to the patch, wherein the depth information represents depths of the points in a direction perpendicular to the patch plane.

[0006] For example, the patch image corresponding to the patch projected onto a patch plane may depict the points of the point cloud included in the patch in two directions, such as an X and Y direction. The points of the point cloud may be projected onto a patch plane approximately perpendicular to a normal vector, normal to a surface of the point cloud at the location of the patch. Also, for example, the patch image comprising depth information for the set of points included in the patch may depict depth information, such as depth distances in a Z direction. To depict the depth information, the depth patch image may include a parameter that varies in intensity based on the depth of points in the point cloud at a particular location in the patch image. For example, the patch image depicting depth information may have a same shape as the patch image representing points projected onto the patch plane. However, the depth information patch image may be an image comprising image attributes, such as one or more colors, that vary in intensity, wherein the intensity of the one or more image attributes corresponds to a depth of the point cloud at a location in the patch image where the image attribute is displayed in the patch image depicting depth. For example, points that are closer to the patch plane may be encoded as darker values in the patch image depicting depth and points that are further away from the patch plane may be encoded as brighter values in the patch image depicting depth, for example in a monochromatic patch image depicting depth. Thus, the depth information patch image when aligned with other patch images representing points projected onto the patch plane may indicate the relative depths of the points projected onto the patch plane, based on respective image attribute intensities at locations in the depth patch image that correspond to locations of the points in the other patch images comprising point cloud points projected onto the patch plane.

[0007] In some embodiments, points of a point cloud may be in a same or nearly same location when projected onto a patch plane. For example, the point cloud might have a depth such that some points are in the same location relative to the patch plane, but at different depths. In such embodiments, multiple patches may be generated for different layers of the point cloud. In some embodiments, subsequent layered patches may encode differences between a previous layer, such that the subsequent layers do not repeat the full amount of data encoded in the previous layer(s). Thus, subsequent layers may have significantly smaller sizes than initial layers.

[0008] The encoder is further configured to pack generated patch images (including a depth patch image and, optionally, one or more additional patch images for one or more other attributes) for each of the determined patches into one or more image frames and encode the one or more image frames. In some embodiments, the encoder may utilize various image or video encoding techniques to encode the one or more image frames. For example, the encoder may utilize a video encoder in accordance with the High Efficiency Video Coding (HEVC/H.265) standard or other suitable standards such as, the Advanced Video Coding (AVC/H.265) standard, the AOMedia Video 1 (AV1) video coding format produced by the Alliance for Open Media (AOM), etc. In some embodiments, the encoder may utilize an image encoder in accordance with a Motion Picture Experts Group (MPEG), a Joint Photography Experts Group (JPEG) standard, an International Telecommunication Union-Telecommunication standard (e.g. ITU-T standard), etc.

[0009] In some embodiments, colors of patch images packed into image frames may be converted into a different color space or may be sub-sampled to further compress the image frames. For example, in some embodiments a 4:4:4 R’G’B’ color space may be converted into a 4:2:0 YCbCr color space. Additionally, a color conversion process may determine an optimal luma value and corresponding chroma values. For example, a an optimal luma value may be selected that reduces a converted size of the fame image while minimizing distortion of the decompressed point cloud colors as compared to the original non-compressed point cloud. In some embodiments, an iterative approach may be used to determine an optimal luma value. In other embodiments, one or more optimization equations may be applied to determine an optimal luma and corresponding chroma values.

[0010] Such a system may further account for distortion caused by projecting the point cloud onto patches and packing the patches into image frames. Additionally, such a system may account for distortion caused by video encoding and decoding the image frames comprising packed patches. To do this, a closed-loop color conversion module may take as an input a reference point cloud original color and a video compressed image frame comprising packed patches, wherein the packed patches of the image frame have been converted from a first color space to a second color space. The closed-loop color conversion module may decompress the compressed image frame using a video decoder and furthermore reconstruct the point cloud using the decompressed image frames. The closed-loop color conversion module may then determine color values for points of the decompressed point cloud based on attribute and/or texture information included in the decompressed patches of the decompressed image frames (in the converted color space). The closed-loop color conversion module may then compare the down sampled and up sampled colors of the reconstructed point cloud to the colors of the original non-compressed point cloud. Based on this comparison, the closed-loop color conversion module may then adjust one or more parameters used to convert the image frames from the original color space to the second color space, wherein the one or more parameters are adjusted to improve quality of the final decompressed point cloud colors and to reduce a size of the compressed point cloud.

[0011] In some embodiments, a decoder is configured to receive one or more encoded image frames comprising patch images for a plurality of patches of a compressed point cloud, wherein, for each patch, the one or more encoded image frames comprise: a patch image comprising a set of points of the patch projected onto a patch plane and a patch image comprising depth information for the set of points of the patch, wherein the depth information indicates depths of the points of the patch in a direction perpendicular to the patch plane. In some embodiments, a depth patch image may be packed into an image frame with other attribute patch images. For example, a decoder may receive one or more image frames comprising packed patch images as generated by the encoder described above.

[0012] The decoder is further configured to decode the one or more encoded image frames comprising the patch images. In some embodiments, the decoder may utilize a video decoder in accordance with the High Efficiency Video Coding (HEVC) standard or other suitable standards such as, the Advanced Video Coding (AVC) standard, the AOMedia Video 1 (AV1) video coding format, etc. In some embodiments, the decoder may utilize an image decoder in accordance with a Motion Picture Experts Group (MPEG) or a Joint Photography Experts Group (JPEG) standard, etc.

[0013] The decoder is further configured to determine, for each patch, spatial information for the set of points of the patch based, at least in part, on the patch image comprising the set of points of the patch projected onto the patch plane and the patch image comprising the depth information for the set of points of the patch, and generate a decompressed version of the compressed point cloud based, at least in part, on the determined spatial information for the plurality of patches and the attribute information included in the patches.

[0014] In some embodiments, a method includes receiving one or more encoded image frames comprising patch images for a plurality of patches of a compressed point cloud, wherein, for each patch, the one or more encoded image frames comprise: a patch image comprising a set of points of the patch projected onto a patch plane and a patch image comprising depth information for the set of points of the patch, wherein the depth information indicates depths of the points of the patch in a direction perpendicular to the patch plane. The method further includes decoding the one or more encoded image frames comprising the patch images. In some embodiments, decoding may be performed in accordance with the High Efficiency Video Coding (HEVC) standard or other suitable standards such as, the Advanced Video Coding (AVC) standard, an AOMedia Video 1 (AV1) video coding format, etc. In some embodiments, decoding may be performed in accordance with a Motion Picture Experts Group (MPEG) or a Joint Photography Experts Group (JPEG) standard, etc.

[0015] The method further includes determining, for each patch, spatial information for the set of points of the patch based, at least in part, on the patch image comprising the set of points of the patch projected onto the patch plane and the patch image comprising the depth information for the set of points of the patch and generating a decompressed version of the compressed point cloud based, at least in part, on the determined spatial information for the plurality of patches.

[0016] In some embodiments, a non-transitory computer-readable medium stores program instructions that, when executed by one or more processors, cause the one or more processors to implement an encoder as described herein to compress attribute information of a point cloud.

[0017] In some embodiments, a non-transitory computer-readable medium stores program instructions that, when executed by one or more processors, cause the one or more processors to implement a decoder as described herein to decompress attribute information of a point cloud.

BRIEF DESCRIPTION OF THE DRAWINGS

[0018] FIG. 1 illustrates a system comprising a sensor that captures information for points of a point cloud and an encoder that compresses spatial information and attribute information of the point cloud, where the compressed spatial and attribute information is sent to a decoder, according to some embodiments.

[0019] FIG. 2A illustrates components of an encoder for encoding intra point cloud frames, according to some embodiments.

[0020] FIG. 2B illustrates components of a decoder for decoding intra point cloud frames, according to some embodiments.

[0021] FIG. 2C illustrates components of an encoder for encoding inter point cloud frames, according to some embodiments.

[0022] FIG. 2D illustrates components of a decoder for decoding inter point cloud frames, according to some embodiments.

[0023] FIG. 2E illustrates components of a closed-loop color conversion module, according to some embodiments.

[0024] FIG. 2F illustrates an example process for determining a quality metric for a point cloud upon which an operation has been performed, according to some embodiments.

[0025] FIG. 3A illustrates an example patch segmentation process, according to some embodiments.

[0026] FIG. 3B illustrates an example image frame comprising packed patch images and padded portions, according to some embodiments.

[0027] FIG. 3C illustrates an example image frame comprising patch portions and padded portions, according to some embodiments.

[0028] FIG. 3D illustrates a point cloud being projected onto multiple projections, according to some embodiments.

[0029] FIG. 3E illustrates a point cloud being projected onto multiple parallel projections, according to some embodiments.

[0030] FIG. 4A illustrates a process for compressing attribute and spatial information of a point cloud, according to some embodiments.

[0031] FIG. 4B illustrates a process for decompressing attribute and spatial information of a point cloud, according to some embodiments.

[0032] FIG. 4C illustrates patch images being generated and packed into an image frame to compress attribute and spatial information of a point cloud, according to some embodiments.

[0033] FIG. 4D illustrates patch images being generated and packed into an image frame to compress attribute and spatial information of a moving or changing point cloud, according to some embodiments.

[0034] FIG. 4E illustrates a decoder receiving image frames comprising patch images, patch information, and an occupancy map, and generating a decompressed representation of a point cloud, according to some embodiments.

[0035] FIG. 4F illustrates an encoder, adjusting encoding based on one or more masks for a point cloud, according to some embodiments.

[0036] FIG. 4G illustrates a decoder, adjusting decoding based on one or more masks for a point cloud, according to some embodiments.

[0037] FIG. 4H illustrates more detail regarding compression of an occupancy map, according to some embodiments.

[0038] FIG. 4I illustrates example blocks and traversal patterns for compressing an occupancy map, according to some embodiments.

[0039] FIG. 5 illustrates compressed point cloud information being used in a 3-D telepresence application, according to some embodiments.

[0040] FIG. 6 illustrates compressed point cloud information being used in a virtual reality application, according to some embodiments.

[0041] FIG. 7 illustrates an example computer system that may implement an encoder or decoder, according to some embodiments.

[0042] This specification includes references to “one embodiment” or “an embodiment.” The appearances of the phrases “in one embodiment” or “in an embodiment” do not necessarily refer to the same embodiment. Particular features, structures, or characteristics may be combined in any suitable manner consistent with this disclosure.

[0043] “Comprising.” This term is open-ended. As used in the appended claims, this term does not foreclose additional structure or steps. Consider a claim that recites: “An apparatus comprising one or more processor units … .” Such a claim does not foreclose the apparatus from including additional components (e.g., a network interface unit, graphics circuitry, etc.).

[0044] “Configured To.” Various units, circuits, or other components may be described or claimed as “configured to” perform a task or tasks. In such contexts, “configured to” is used to connote structure by indicating that the units/circuits/components include structure (e.g., circuitry) that performs those task or tasks during operation. As such, the unit/circuit/component can be said to be configured to perform the task even when the specified unit/circuit/component is not currently operational (e.g., is not on). The units/circuits/components used with the “configured to” language include hardware–for example, circuits, memory storing program instructions executable to implement the operation, etc. Reciting that a unit/circuit/component is “configured to” perform one or more tasks is expressly intended not to invoke 35 U.S.C. .sctn. 112(f), for that unit/circuit/component. Additionally, “configured to” can include generic structure (e.g., generic circuitry) that is manipulated by software and/or firmware (e.g., an FPGA or a general-purpose processor executing software) to operate in manner that is capable of performing the task(s) at issue. “Configure to” may also include adapting a manufacturing process (e.g., a semiconductor fabrication facility) to fabricate devices (e.g., integrated circuits) that are adapted to implement or perform one or more tasks.

[0045] “First,” “Second,” etc. As used herein, these terms are used as labels for nouns that they precede, and do not imply any type of ordering (e.g., spatial, temporal, logical, etc.). For example, a buffer circuit may be described herein as performing write operations for “first” and “second” values. The terms “first” and “second” do not necessarily imply that the first value must be written before the second value.

[0046] “Based On.” As used herein, this term is used to describe one or more factors that affect a determination. This term does not foreclose additional factors that may affect a determination. That is, a determination may be solely based on those factors or based, at least in part, on those factors. Consider the phrase “determine A based on B.” While in this case, B is a factor that affects the determination of A, such a phrase does not foreclose the determination of A from also being based on C. In other instances, A may be determined based solely on B.

DETAILED DESCRIPTION

[0047] As data acquisition and display technologies have become more advanced, the ability to capture point clouds comprising thousands or millions of points in 2-D or 3-D space, such as via LIDAR systems, has increased. Also, the development of advanced display technologies, such as virtual reality or augmented reality systems, has increased potential uses for point clouds. However, point cloud files are often very large and may be costly and time-consuming to store and transmit. For example, communication of point clouds over private or public networks, such as the Internet, may require considerable amounts of time and/or network resources, such that some uses of point cloud data, such as real-time uses, may be limited. Also, storage requirements of point cloud files may consume a significant amount of storage capacity of devices storing the point cloud files, which may also limit potential applications for using point cloud data.

[0048] In some embodiments, an encoder may be used to generate a compressed point cloud to reduce costs and time associated with storing and transmitting large point cloud files. In some embodiments, a system may include an encoder that compresses attribute and/or spatial information of a point cloud file such that the point cloud file may be stored and transmitted more quickly than non-compressed point clouds and in a manner that the point cloud file may occupy less storage space than non-compressed point clouds. In some embodiments, compression of attributes of points in a point cloud may enable a point cloud to be communicated over a network in real-time or in near real-time. For example, a system may include a sensor that captures attribute information about points in an environment where the sensor is located, wherein the captured points and corresponding attributes make up a point cloud. The system may also include an encoder that compresses the captured point cloud attribute information. The compressed attribute information of the point cloud may be sent over a network in real-time or near real-time to a decoder that decompresses the compressed attribute information of the point cloud. The decompressed point cloud may be further processed, for example to make a control decision based on the surrounding environment at the location of the sensor. The control decision may then be communicated back to a device at or near the location of the sensor, wherein the device receiving the control decision implements the control decision in real-time or near real-time. In some embodiments, the decoder may be associated with an augmented reality system and the decompressed attribute information may be displayed or otherwise used by the augmented reality system. In some embodiments, compressed attribute information for a point cloud may be sent with compressed spatial information for points of the point cloud. In other embodiments, spatial information and attribute information may be separately encoded and/or separately transmitted to a decoder.

[0049] In some embodiments, a system may include a decoder that receives one or more sets of point cloud data comprising compressed attribute information via a network from a remote server or other storage device that stores the one or more point cloud files. For example, a 3-D display, a holographic display, or a head-mounted display may be manipulated in real-time or near real-time to show different portions of a virtual world represented by point clouds. In order to update the 3-D display, the holographic display, or the head-mounted display, a system associated with the decoder may request point cloud data from the remote server based on user manipulations of the displays, and the point cloud data may be transmitted from the remote server to the decoder and decoded by the decoder in real-time or near real-time. The displays may then be updated with updated point cloud data responsive to the user manipulations, such as updated point attributes.

[0050] In some embodiments, a system, may include one or more LIDAR systems, 3-D cameras, 3-D scanners, etc., and such sensor devices may capture spatial information, such as X, Y, and Z coordinates for points in a view of the sensor devices. In some embodiments, the spatial information may be relative to a local coordinate system or may be relative to a global coordinate system (for example, a Cartesian coordinate system may have a fixed reference point, such as a fixed point on the earth, or may have a non-fixed local reference point, such as a sensor location).

[0051] In some embodiments, such sensors may also capture attribute information for one or more points, such as color attributes, reflectivity attributes, velocity attributes, acceleration attributes, time attributes, modalities, and/or various other attributes. In some embodiments, other sensors, in addition to LIDAR systems, 3-D cameras, 3-D scanners, etc., may capture attribute information to be included in a point cloud. For example, in some embodiments, a gyroscope or accelerometer, may capture motion information to be included in a point cloud as an attribute associated with one or more points of the point cloud. For example, a vehicle equipped with a LIDAR system, a 3-D camera, or a 3-D scanner may include the vehicle’s direction and speed in a point cloud captured by the LIDAR system, the 3-D camera, or the 3-D scanner. For example, when points in a view of the vehicle are captured they may be included in a point cloud, wherein the point cloud includes the captured points and associated motion information corresponding to a state of the vehicle when the points were captured.

Example System Arrangement

[0052] FIG. 1 illustrates a system comprising a sensor that captures information for points of a point cloud and an encoder that compresses attribute information of the point cloud, where the compressed attribute information is sent to a decoder, according to some embodiments.

[0053] System 100 includes sensor 102 and encoder 104. Sensor 102 captures a point cloud 110 comprising points representing structure 106 in view 108 of sensor 102. For example, in some embodiments, structure 106 may be a mountain range, a building, a sign, an environment surrounding a street, or any other type of structure. In some embodiments, a captured point cloud, such as captured point cloud 110, may include spatial and attribute information for the points included in the point cloud. For example, point A of captured point cloud 110 comprises X, Y, Z coordinates and attributes 1, 2, and 3. In some embodiments, attributes of a point may include attributes such as R, G, B color values, a velocity at the point, an acceleration at the point, a reflectance of the structure at the point, a time stamp indicating when the point was captured, a string-value indicating a modality when the point was captured, for example “walking”, or other attributes. The captured point cloud 110 may be provided to encoder 104, wherein encoder 104 generates a compressed version of the point cloud (compressed attribute information 112) that is transmitted via network 114 to decoder 116. In some embodiments, a compressed version of the point cloud, such as compressed attribute information 112, may be included in a common compressed point cloud that also includes compressed spatial information for the points of the point cloud or, in some embodiments, compressed spatial information and compressed attribute information may be communicated as separate sets of data.

[0054] In some embodiments, encoder 104 may be integrated with sensor 102. For example, encoder 104 may be implemented in hardware or software included in a sensor device, such as sensor 102. In other embodiments, encoder 104 may be implemented on a separate computing device that is proximate to sensor 102.

Example Intra-Frame Encoder

[0055] FIG. 2A illustrates components of an encoder for encoding intra point cloud frames, according to some embodiments. In some embodiments, the encoder described above in regard to FIG. 1 may operate in a similar manner as encoder 200 described in FIG. 2A and encoder 250 described in FIG. 2C.

[0056] The encoder 200 receives uncompressed point cloud 202 and generates compressed point cloud information 204. In some embodiments, an encoder, such as encoder 200, may receive the uncompressed point cloud 202 from a sensor, such as sensor 102 illustrated in FIG. 1, or, in some embodiments, may receive the uncompressed point cloud 202 from another source, such as a graphics generation component that generates the uncompressed point cloud in software, as an example.

[0057] In some embodiments, an encoder, such as encoder 200, includes decomposition into patches module 206, packing module 208, spatial image generation module 210, texture image generation module 212, and attribute information generation module 214. In some embodiments, an encoder, such as encoder 200, also includes image frame padding module 216, closed loop color conversion module 217, video compression module 218 and multiplexer 224. In addition, in some embodiments an encoder, such as encoder 200, may include an occupancy map compression module, such as occupancy map compression module 220, and an auxiliary patch information compression module, such as auxiliary patch information compression module 222. In some embodiments, an encoder, such as encoder 200, converts a 3D point cloud into an image-based representation along with some meta data (e.g., occupancy map and patch info) necessary to convert the compressed point cloud back into a decompressed point cloud.

[0058] In some embodiments, the conversion process decomposes the point cloud into a set of patches (e.g., a patch is defined as a contiguous subset of the surface described by the point cloud), which may be overlapping or not, such that each patch may be described by a depth field with respect to a plane in 2D space. More details about the patch decomposition process are provided below with regard to FIGS. 3A-3C.

[0059] After or in conjunction with the patches being determined for the point cloud being compressed, a 2D sampling process is performed in planes associated with the patches. The 2D sampling process may be applied in order to approximate each patch with a uniformly sampled point cloud, which may be stored as a set of 2D patch images describing the geometry/texture/attributes of the point cloud at the patch location. The “Packing” module 208 may store the 2D patch images associated with the patches in a single (or multiple) 2D images, referred to herein as “image frames.” In some embodiments, a packing module, such as packing module 208, may pack the 2D patch images such that the packed 2D patch images do not overlap (even though an outer bounding box for one patch image may overlap an outer bounding box for another patch image). Also, the packing module may pack the 2D patch images in a way that minimizes non-used images pixels of the image frame.

[0060] In some embodiments, “Geometry/Texture/Attribute generation” modules, such as modules 210, 212, and 214, generate 2D patch images associated with the geometry/texture/attributes, respectively, of the point cloud at a given patch location. As noted before, a packing process, such as performed by packing module 208, may leave some empty spaces between 2D patch images packed in an image frame. Also, a padding module, such as image frame padding module 216, may fill in such areas in order to generate an image frame that may be suited for 2D video and image codecs.

[0061] In some embodiments, an occupancy map (e.g., binary information describing for each pixel or block of pixels whether the pixel or block of pixels are padded or not) may be generated and compressed, for example by occupancy map compression module 220. The occupancy map may be sent to a decoder to enable the decoder to distinguish between padded and non-padded pixels of an image frame.

[0062] Note that other metadata associated with patches may also be sent to a decoder for use in the decompression process. For example, patch information indicating sizes and shapes of patches determined for the point cloud and packed in an image frame may be generated and/or encoded by an auxiliary patch-information compression module, such as auxiliary patch-information compression module 222. In some embodiments one or more image frames may be encoded by a video encoder, such as video compression module 218. In some embodiments, a video encoder, such as video compression module 218, may operate in accordance with the High Efficiency Video Coding (HEVC) standard or other suitable video encoding standard. In some embodiments, encoded video images, encoded occupancy map information, and encoded auxiliary patch information may be multiplexed by a multiplexer, such as multiplexer 224, and provided to a recipient as compressed point cloud information, such as compressed point cloud information 204.

[0063] In some embodiments, an occupancy map may be encoded and decoded by a video compression module, such as video compression module 218. This may be done at an encoder, such as encoder 200, such that the encoder has an accurate representation of what the occupancy map will look like when decoded by a decoder. Also, variations in image frames due to lossy compression and decompression may be accounted for by an occupancy map compression module, such as occupancy map compression module 220, when determining an occupancy map for an image frame. In some embodiments, various techniques may be used to further compress an occupancy map, such as described in FIGS. 4H-4I.

Example Intra-Frame Decoder

[0064] FIG. 2B illustrates components of a decoder for decoding intra point cloud frames, according to some embodiments. Decoder 230 receives compressed point cloud information 204, which may be the same compressed point cloud information 204 generated by encoder 200. Decoder 230 generates reconstructed point cloud 246 based on receiving the compressed point cloud information 204. In some embodiments, a decoder, such as decoder 230, includes a de-multiplexer 232, a video decompression module 234, an occupancy map decompression module 236, and an auxiliary patch-information decompression module 238. Additionally a decoder, such as decoder 230 includes a point cloud generation module 240, which reconstructs a point cloud based on patch images included in one or more image frames included in the received compressed point cloud information, such as compressed point cloud information 204. In some embodiments, a decoder, such as decoder 203, further comprises a smoothing filter, such as smoothing filter 244. In some embodiments, a smoothing filter may smooth incongruences at edges of patches, wherein data included in patch images for the patches has been used by the point cloud generation module to recreate a point cloud from the patch images for the patches. In some embodiments, a smoothing filter may be applied to the pixels located on the patch boundaries to alleviate the distortions that may be caused by the compression/decompression process.

Example Inter-Frame Encoder

[0065] FIG. 2C illustrates components of an encoder for encoding inter point cloud frames, according to some embodiments. An inter point cloud encoder, such as inter point cloud encoder 250, may encode an image frame, while considering one or more previously encoded/decoded image frames as references.

[0066] In some embodiments, an encoder for inter point cloud frames, such as encoder 250, includes a point cloud re-sampling module 252, a 3-D motion compensation and delta vector prediction module 254, a spatial image generation module 256, a texture image generation module 258, and an attribute image generation module 260. In some embodiments, an encoder for inter point cloud frames, such as encoder 250, may also include an image padding module 262, a closed-loop color conversion module 263, and a video compression module 264. An encoder for inter point cloud frames, such as encoder 250, may generate compressed point cloud information, such as compressed point cloud information 266. In some embodiments, the compressed point cloud information may reference point cloud information previously encoded by the encoder, such as information from or derived from one or more reference image frames. In this way an encoder for inter point cloud frames, such as encoder 250, may generate more compact compressed point cloud information by not repeating information included in a reference image frame, and instead communicating differences between the reference frames and a current state of the point cloud.

[0067] In some embodiments, an encoder, such as encoder 250, may be combined with or share modules with an intra point cloud frame encoder, such as encoder 200. In some embodiments, a point cloud re-sampling module, such as point cloud re-sampling module 252, may resample points in an input point cloud image frame in order to determine a one-to-one mapping between points in patches of the current image frame and points in patches of a reference image frame for the point cloud. In some embodiments, a 3D motion compensation & delta vector prediction module, such as a 3D motion compensation & delta vector prediction module 254, may apply a temporal prediction to the geometry/texture/attributes of the resampled points of the patches. The prediction residuals may be stored into images, which may be padded and compressed by using video/image codecs. In regard to spatial changes for points of the patches between the reference frame and a current frame, a 3D motion compensation & delta vector prediction module 254, may determine respective vectors for each of the points indicating how the points moved from the reference frame to the current frame. A 3D motion compensation & delta vector prediction module 254, may then encode the motion vectors using different image parameters. For example, changes in the X direction for a point may be represented by an amount of red included at the point in a patch image that includes the point. In a similar manner, changes in the Y direction for a point may be represented by an amount of blue included at the point in a patch image that includes the point. Also, in a similar manner, changes in the Z direction for a point may be represented by an amount of green included at the point in a patch image that includes the point. In some embodiments, other characteristics of an image included in a patch image may be adjusted to indicate motion of points included in the patch between a reference frame for the patch and a current frame for the patch.

Example Inter-Frame Decoder

[0068] FIG. 2D illustrates components of a decoder for decoding inter point cloud frames, according to some embodiments. In some embodiments, a decoder, such as decoder 280, includes a video decompression module 270, an inverse 3D motion compensation and inverse delta prediction module 272, a point cloud generation module 274, and a smoothing filter 276. In some embodiments, a decoder, such as decoder 280 may be combined with a decoder, such as decoder 230, or may share some components with the decoder, such as a video decompression module and/or smoothing filter. In decoder 280, the video/image streams are first decoded, then an inverse motion compensation and delta prediction procedure may be applied. The obtained images are then used in order to reconstruct a point cloud, which may be smoothed as described previously to generate a reconstructed point cloud 282.

……
……
……