阅读说明：本技术 用于点云压缩的量化步长参数 (Quantization step size parameter for point cloud compression ) 是由 A·扎格托 D·格拉兹斯 A·塔巴塔拜 于 2020-03-09 设计创作，主要内容包括：本公开涉及用于点云压缩的量化步长参数。本文中描述了一种对点云压缩算法的属性编码的量化方案进行参数化的技术。基于定点算术,在给定用户输入的量化参数(QP)的情况下,该算法在定点计数法中计算量化步长(QS)。(The present disclosure relates to quantization step size parameters for point cloud compression. Described herein is a technique to parameterize a quantization scheme for attribute coding of a point cloud compression algorithm. Based on fixed point arithmetic, the algorithm calculates the quantization step size (QS) in a fixed point counting method given the user input Quantization Parameter (QP).)

1. A method programmed in a non-transitory memory of a device, comprising:

acquiring three-dimensional content;

implementing geometric encoding on the three-dimensional content; and

fixed-point property encoding is implemented using quantization parameters.

2. The method of claim 1, wherein the quantization parameter is the same for chroma and luma.

3. The method of claim 1, wherein the quantization parameter is different for chrominance and luminance.

4. The method of claim 3, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

5. The method of claim 3, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

6. The method of claim 1, wherein the quantization parameter corresponds to a quantization step size.

7. The method of claim 6, wherein the correspondence between quantization parameters and quantization steps is made using a table.

8. An apparatus, comprising:

a non-transitory memory for storing an application for:

acquiring three-dimensional content;

implementing geometric encoding on the three-dimensional content; and

implementing fixed point attribute coding by using the quantization parameter; and

a processor coupled to the memory, the processor configured to process the application.

9. The apparatus of claim 8, wherein the quantization parameter is the same for chroma and luma.

10. The apparatus of claim 8, wherein the quantization parameter is different for chroma and luma.

11. The apparatus of claim 10, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

12. The apparatus of claim 10, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

13. The apparatus of claim 8, wherein the quantization parameter corresponds to a quantization step size.

14. The apparatus of claim 13, wherein a correspondence between quantization parameters and quantization steps is made using a table.

15. A system, comprising:

one or more devices for obtaining three-dimensional content; and

an encoder for encoding three-dimensional content by:

implementing geometric encoding on the three-dimensional content; and

fixed-point property encoding is implemented using quantization parameters.

16. The system of claim 15, wherein the quantization parameter is the same for chrominance and luminance.

17. The system of claim 15, wherein the quantization parameter is different for chrominance and luminance.

18. The system of claim 17, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

19. The system of claim 17, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

20. The system of claim 15, wherein the quantization parameter corresponds to a quantization step size.

21. The system of claim 20, wherein the correspondence between quantization parameters and quantization steps is made using a table.

Technical Field

The present invention relates to three-dimensional graphics. More particularly, the present invention relates to encoding of three-dimensional graphics.

Background

Point clouds are considered as candidate formats for transmission of 3D data, which is captured by a 3D scanner, a LIDAR sensor, or used in popular applications such as virtual reality/augmented reality (VR/AR). The point cloud is a collection of points in 3D space. In addition to the spatial position (X, Y, Z), each point typically has associated attributes such as color (R, G, B) or even reflectivity and time stamp (e.g. in LIDAR images). To obtain a high fidelity representation of the target 3D object, the device captures a point cloud of points on the order of thousands or even millions. Furthermore, for dynamic 3D scenes used in VR/AR applications, each frame often has a unique dense point cloud, resulting in the transmission of millions of point clouds per second. For feasible transmission of such large amounts of data, compression is often applied.

In 2017, MPEG issued a proposed levee (CfP) for point cloud compression. After evaluating several proposals, MPEG is considering two different techniques for point cloud compression: 3D native coding techniques (based on octree and similar coding methods), or 3D to 2D projection followed by conventional video coding. In the case of dynamic 3D scenes, MPEG is using test model software based on patch surface modeling (TMC2), patch projection from 3D to 2D images, and encoding of 2D images with a video encoder such as HEVC. This approach has proven to be more efficient than native 3D coding and can achieve competitive bit rates with acceptable quality.

When encoding the point cloud, TMC2 encodes the associated auxiliary information for the tile projection, such as the tile location and bounding box size in the 2D canvas image. For temporal encoding of the side information, patch matches between patches from the current point cloud and patches from the immediately decoded point cloud are used for prediction. The procedure is restricted to the immediate neighborhood and includes applying delta encoding for all frames in the sequence.

Disclosure of Invention

Described herein is a technique to parameterize a quantization scheme for attribute coding of a point cloud compression algorithm. Based on fixed point arithmetic, the algorithm calculates the quantization step size (QS) in a fixed point counting method given the user input Quantization Parameter (QP).

In one aspect, a method programmed in a non-transitory memory of a device includes: obtaining three-dimensional content, implementing geometric structure coding on the three-dimensional content, and implementing fixed-point attribute coding by using quantization parameters. The quantization parameters are the same for chrominance and luminance. The quantization parameters are different for chrominance and luminance. The chrominance quantization parameter is proportional to the luminance quantization parameter. The chrominance quantization parameter is a function of the luminance quantization parameter. The quantization parameter corresponds to a quantization step. The correspondence between quantization parameters and quantization steps is made using a table.

In another aspect, an apparatus comprises: a non-transitory memory for storing an application for: acquiring three-dimensional content, implementing geometric structure coding on the three-dimensional content, and implementing fixed-point attribute coding by using quantization parameters; and a processor coupled to the memory, the processor configured to process the application. The quantization parameters are the same for chrominance and luminance. The quantization parameters are different for chrominance and luminance. The chrominance quantization parameter is proportional to the luminance quantization parameter. The chrominance quantization parameter is a function of the luminance quantization parameter. The quantization parameter corresponds to a quantization step. The correspondence between quantization parameters and quantization steps is made using a table.

In another aspect, a system includes one or more devices for obtaining three-dimensional content, and an encoder for encoding the three-dimensional content by: geometric encoding is performed on the three-dimensional content, and fixed-point attribute encoding is performed using quantization parameters. The quantization parameters are the same for chrominance and luminance. The quantization parameters are different for chrominance and luminance. The chrominance quantization parameter is proportional to the luminance quantization parameter. The chrominance quantization parameter is a function of the luminance quantization parameter. The quantization parameter corresponds to a quantization step. The correspondence between quantization parameters and quantization steps is made using a table.

Drawings

FIG. 1 illustrates a flow diagram of an attribute encoding method according to some embodiments.

Fig. 2 and 3 illustrate the results of a Region Adaptive Hierarchical Transform (RAHT) with QP, according to some embodiments.

Fig. 4 illustrates a block diagram of an exemplary computing device configured to implement the QP point cloud compression method, in accordance with some embodiments.

Detailed Description

Described herein is a technique to parameterize a quantization scheme for attribute coding of a point cloud compression algorithm. Based on fixed point arithmetic, the algorithm calculates the Quantization Step (QS) in a fixed point counting method (normal) given the Quantization Parameter (QP) input by the user.

MPEG (moving picture experts group) is currently defining a standard for Point Cloud Compression (PCC). Point clouds are used to represent three-dimensional scenes and objects and are composed of volume elements (voxels) described by their geometric or appearance properties. TMC13 is a test model software maintained and distributed by MPEG that constantly incorporates new proposals licensed by its contributors. The standard geometry attribute-based compression scheme (referred to as G-PCC) can use a quantization framework to implement attribute coding and is also implemented in TMC13 software. Quantization is the process of mapping a range of values to a single value, resulting in a lossy compression scheme. In the context of G-PCC, quantization reduces the dynamic range of the transformed attribute coefficients.

3D data compression using 3D native techniques is typically done using a mesh or spatial coordinates of the points forming the point cloud. While either approach may be used to represent 3D content, in some applications a point cloud is preferred. In this latter case, geometry compression can be successfully achieved through octrees. In addition to geometry, properties such as color and reflectivity may be compressed. In this case, transforms such as the Region Adaptive Hierarchical Transform (RAHT) and the lifting transform have proven to be the most successful approaches. The most frequent application of this compression architecture envisaged in the test model software TMC13 of MPEG relates to static point clouds and dynamically acquired point clouds.

Only QS was previously used for quantization purposes. Mapping QP to QS using fixed point arithmetic is described herein. Furthermore, by mapping QP to QS using fixed point arithmetic, it is possible to implement a QP-based RDO (rate distortion optimization) framework. Furthermore, by selecting a different QP for each channel, e.g., chroma QP may be different from luma QP, improved bit allocation may be implemented between attribute channels.

QP allows for finer rate-distortion allocation and utilization of fixed point arithmetic. Different quantization parameters may be used for luma and chroma, or a relationship between QPs for different attribute channels may be generated.

FIG. 1 illustrates a flow diagram of an attribute encoding method according to some embodiments. The point cloud data typically contains geometric and color attribute information. In step 100, geometric encoding of three-dimensional content (e.g., point clouds) is performed. The geometry coding can be implemented in any way, such as 3D native coding techniques (based on octree and similar coding methods), or 3D to 2D projection followed by conventional video coding. In step 102, attribute transfer is performed. In step 104, fixed point attribute encoding is performed. For color and attribute coding, a graphical transformation can be used. Fixed point attribute coding sends QP instead of QS in the bitstream. In some embodiments, the same QP is used for luma and chroma, in some embodiments different QPs are used. Improved bit allocation is allowed by using different QPs for chroma and luma. QP is sent instead of QS in the TMC13 bitstream. The updated configuration file uses a QP instead of QSLuma or QSChroma. In step 106, the bit stream is encoded. After geometry coding and fixed-point property coding, a bitstream is generated and sent to a decoder. In step 108, geometry decoding is performed. Geometry decoding is configured based on geometry encoding to ensure correct decoding. In step 110, fixed point attribute decoding is performed. Fixed-point attribute decoding utilizes QPs sent in the bitstream. In some embodiments, fewer or additional steps are performed. In some embodiments, the order of the steps is modified.

The following table illustrates the correspondence between QS and QP under CTC conditions:

rate of speed	quantizationStepLuma	quantQPFactor
			r01	256	52
r02	128	46
			r03	64	40
r04	32	34
			r05	16	28
r06	8	22

Obtaining a quantization step size from a quantization parameter QP using equation 1_Δstep。

Using fixed-point counting (ratio k), equation 1 becomes

Wherein

Assuming α -4 and k-8, the following quantization is obtained:

Δ₀＝{161,181,203,228,256,287} (4)

once the quantization of the transformed attribute is implemented by QP, the lagrangian rate-distortion function J can be defined as:

J(QP)＝R(QP)+λ(QP)D(QP)

the bitrate distribution between luminance and chrominance is controlled by the QP.

Considering that luminance and chrominance have specific characteristics, a different QP may be used for each channel to achieve better bit allocation. QPL (QP for luminance) and QPC (QP for chrominance) are quantization parameters for luminance and chrominance, respectively. Thus, QPC as a function of QPL is as follows:

QPC＝φ(QPL)。

considering that different attributes have specific characteristics, different QPs can be used for each attribute to achieve better bit allocation. QP may be a base quantization parameter, QPA_0..n-1May be for attribute A₀And A_n-1The quantization parameter of (1). QPA is formed_iWrite the function of QP:

QPA_i＝ψ_i(QP)，i＝0..n-1。

fig. 2 and 3 show results of RAHT with QP according to some embodiments.

Fig. 4 illustrates a block diagram of an exemplary computing device configured to implement the QP point cloud compression method, in accordance with some embodiments. Computing device 400 can be used to acquire, store, calculate, process, communicate, and/or display information, such as images and videos, including 3D content. Computing device 400 can implement any aspect of point cloud compression. In general, a hardware architecture suitable for implementing the computing device 400 includes a network interface 402, memory 404, a processor 406, I/O device(s) 408, a bus 410, and a storage device 412. The choice of processor is not critical as long as the appropriate processor with sufficient speed is selected. The memory 404 may be any conventional computer memory known in the art. The storage device 412 may comprise a hard disk drive, CDROM, CDRW, DVD, DVDRW, high definition disc/drive, ultra high definition drive, flash memory card, or any other storage device. Computing device 400 may include one or more network interfaces 402. One example of a network interface includes a network card connected to an ethernet or other type of LAN. I/O device(s) 408 may include one or more of the following: keyboard, mouse, monitor, screen, printer, modem, touch screen, button interface, and other devices. The QP point cloud compression application(s) 430 used to implement the QP point cloud compression method may be stored in the storage 412 and memory 404 and processed as the application is typically processed. More or fewer components than are shown in fig. 4 may be included in computing device 400. In some embodiments, QP point cloud compression hardware 420 is included. Although the computing device 400 in fig. 4 includes the application 430 and hardware 420 for the QP point cloud compression method, the QP point cloud compression method can be implemented on the computing device by hardware, firmware, software, or any combination thereof. For example, in some embodiments, the QP point cloud compression application 430 is programmed in memory and executed using a processor. In another example, in some embodiments, the QP point cloud compression hardware 420 is programmed hardware logic including gates specifically designed to implement the QP point cloud compression method.

In some embodiments, the QP point cloud compression application(s) 430 includes several applications and/or modules. In some embodiments, the module further comprises one or more sub-modules. In some embodiments, fewer or additional modules may be included.

In some embodiments, the QP point cloud compression hardware 420 includes camera components such as a lens, an image sensor, and/or any other camera component.

Examples of suitable computing devices include personal computers, laptop computers, computer workstations, servers, mainframe computers, handheld computers, personal digital assistants, cellular/mobile phones, smart appliances, game consoles, digital cameras, digital camcorders, camera phones, smart phones, portable music players, tablet computers, mobile devices, video players, video disc writers/players (e.g., DVD writer/player, high definition disc writer/player, ultra high definition disc writer/player), televisions, home entertainment systems, augmented reality devices, virtual reality devices, smart jewelry (e.g., smart watches), vehicles (e.g., autonomous vehicles), or any other suitable computing device.

To utilize the QP point cloud compression method, a device acquires or receives 3D content and processes and/or transmits the content in an optimized manner, allowing the 3D content to be displayed correctly and efficiently. The QP point cloud compression method may be implemented with the help of a user or automatically without user involvement.

In operation, in addition to using different QPs for luminance and chrominance or any given set of attributes, the QP point cloud compression method uses QP instead of QS to achieve a finer rate-distortion allocation, allowing a better bit rate allocation.

Some embodiments of quantization step size parameters for point cloud compression

1. A method programmed in a non-transitory memory of a device, comprising:

acquiring three-dimensional content;

implementing geometric encoding on the three-dimensional content; and

fixed-point property encoding is implemented using quantization parameters.

2. The method of clause 1, wherein the quantization parameter is the same for chroma and luma.

3. The method of clause 1, wherein the quantization parameter is different for chroma and luma.

4. The method of clause 3, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

5. The method of clause 3, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

6. The method of clause 1, wherein the quantization parameter corresponds to a quantization step size.

7. The method of clause 6, wherein the correspondence between the quantization parameter and the quantization step is made using a table.

8. An apparatus, comprising:

a non-transitory memory for storing an application for:

acquiring three-dimensional content;

implementing geometric encoding on the three-dimensional content; and

implementing fixed point attribute coding by using the quantization parameter; and

a processor coupled to the memory, the processor configured to process the application.

9. The apparatus of clause 8, wherein the quantization parameter is the same for chroma and luma.

10. The apparatus of clause 8, wherein the quantization parameter is different for chroma and luma.

11. The apparatus of clause 10, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

12. The apparatus of clause 10, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

13. The apparatus of clause 8, wherein the quantization parameter corresponds to a quantization step size.

14. The apparatus of clause 13, wherein the correspondence between quantization parameters and quantization steps is made using a table.

15. A system, comprising:

one or more devices for obtaining three-dimensional content; and

an encoder for encoding three-dimensional content by:

implementing geometric encoding on the three-dimensional content; and

fixed-point property encoding is implemented using quantization parameters.

16. The system of clause 15, wherein the quantization parameter is the same for chroma and luma.

17. The system of clause 15, wherein the quantization parameter is different for chroma and luma.

18. The system of clause 17, wherein the chrominance quantization parameter is proportional to the luminance quantization parameter.

19. The system of clause 17, wherein the chrominance quantization parameter is a function of the luminance quantization parameter.

20. The system of clause 15, wherein the quantization parameter corresponds to a quantization step size.

21. The system of clause 20, wherein the correspondence between the quantization parameter and the quantization step size is made using a table.

The foregoing describes the invention in terms of specific embodiments incorporating details to facilitate the understanding of the principles of construction and operation of the invention. Reference herein to specific embodiments and details thereof is not intended to limit the scope of the claims appended hereto. Those skilled in the art will readily recognize various other modifications that may be made in the embodiment chosen for illustration without departing from the spirit and scope of the invention as defined by the claims.