Search

EP-4391538-B1 - POINT CLOUD CODING METHOD AND APPARATUS, POINT CLOUD DECODING METHOD AND APPARATUS, AND COMPUTER AND STORAGE MEDIUM

EP4391538B1EP 4391538 B1EP4391538 B1EP 4391538B1EP-4391538-B1

Inventors

  • ZHU, WENJIE

Dates

Publication Date
20260506
Application Date
20221208

Claims (14)

  1. A point cloud encoding method, performed by a computer device, characterized in that the method comprises: obtaining encoding limit information, and obtaining a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information (S301) ; and encoding the to-be-encoded transform coefficient sequence to obtain a point cloud group bitstream corresponding to the to-be-encoded point cloud group (S303), wherein the obtaining of a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information comprises: obtaining a group of to-be-encoded point cloud groups from point cloud groups based on the encoding limit information, wherein the encoding limit information comprises a group quantity limit threshold and a quantity of the to-be-encoded point cloud groups comprised in the group is less than or equal to the group quantity limit threshold, performing point cloud prediction processing on the to-be-encoded point cloud group to obtain attribute predicted values corresponding to to-be-encoded point cloud points included in the to-be-encoded point cloud group; obtaining attribute residuals of the to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values; obtaining transform coefficients of to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values and the obtained attribute residuals; and sorting transform coefficients of the to-be-encoded point cloud points included in the to-be-encoded point cloud group to obtain the to-be-encoded transform coefficient sequence.
  2. The method according to claim 1, wherein the encoding limit information further comprises a cache limit threshold and a point cloud group size threshold, and the group quantity limit threshold is obtained based on the cache limit threshold and the point cloud group size threshold.
  3. The method according to claim 1, wherein the group quantity limit threshold is determined based on a grouping manner of point cloud groups, and the grouping manner comprises at least one of the following manners: a grouping manner based on a point cloud group size threshold; a grouping manner based on moving sequences obtained through respectively shifting coordinate codewords of k point cloud points, K being a positive integer; or a grouping manner based on a default group quantity.
  4. The method according to claim 1, wherein the transform coefficients of the to-be-encoded point cloud points comprised in the to-be-encoded point cloud group comprise a first transform coefficient and a second transform coefficient; the to-be-encoded transform coefficient sequence comprises a first coefficient subsequence and a second coefficient subsequence; a transform coefficient comprised in the first coefficient subsequence is the first transform coefficient and is consecutive in the transform coefficient sequence; and a transform coefficient comprised in the second coefficient subsequence is the second transform coefficient and is consecutive in the transform coefficient sequence.
  5. The method according to claim 1, wherein the transform coefficients of the to-be-encoded point cloud points comprised in the to-be-encoded point cloud group comprise a first transform coefficient and a second transform coefficient; a transform coefficient adjacent to the first transform coefficient in the to-be-encoded transform coefficient sequence is the second transform coefficient, and a transform coefficient adjacent to the second transform coefficient in the to-be-encoded transform coefficient sequence is the first transform coefficient.
  6. The method according to claim 1, wherein an attribute parameter of the to-be-encoded point cloud points comprises b attribute components, b being a positive integer; the transform coefficients of the to-be-encoded point cloud points comprise first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, and second transform coefficients corresponding to the b attribute components; and the to-be-encoded transform coefficient sequence comprises coefficient subsequences corresponding to the b attribute components, a coefficient subsequence comprising a first transform coefficient and a second transform coefficient of a corresponding attribute component.
  7. The method according to claim 1, wherein an attribute parameter of the to-be-encoded point cloud points comprises b attribute components, b being a positive integer; the transform coefficients of the to-be-encoded point cloud points comprise first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, and second transform coefficients corresponding to the b attribute components; and the first transform coefficients corresponding to the b attribute components are consecutive in the to-be-encoded transform coefficient sequence, and the second transform coefficients corresponding to the b attribute components are consecutive in the to-be-encoded transform coefficient sequence.
  8. The method according to claim 1, wherein an attribute parameter of the to-be-encoded point cloud points comprises b attribute components, b being a positive integer; the transform coefficients of the to-be-encoded point cloud points comprise first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, and second transform coefficients corresponding to the b attribute components; the to-be-encoded transform coefficient sequence comprises a first hybrid subsequence and a second hybrid subsequence; the first hybrid subsequence comprises the first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points; first transform coefficients of associated attribute components of a same to-be-encoded point cloud point in the first hybrid subsequence are adjacent, the associated attribute components being at least two attribute components whose similarities are greater than a component similarity threshold, or at least two attribute components that are associated by default; and first transform coefficients of remaining attribute components of different to-be-encoded point cloud points in the first hybrid subsequence are consecutive, the remaining attribute components being attribute components other than the associated attribute components in the b attribute components; and the second hybrid subsequence comprises the second transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, and second transform coefficients of the b attribute components of a same to-be-encoded point cloud point in the second hybrid subsequence are adjacent in sequence.
  9. The method according to claim 1, wherein an attribute parameter of the to-be-encoded point cloud points comprises b attribute components, b being a positive integer; the transform coefficients of the to-be-encoded point cloud points comprise first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, and second transform coefficients corresponding to the b attribute components; the to-be-encoded transform coefficient sequence comprises a third hybrid subsequence and a fourth hybrid subsequence; the third hybrid subsequence comprises the first transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points; first transform coefficients of associated attribute components of a same to-be-encoded point cloud point in the third hybrid subsequence are adjacent, the associated attribute components being at least two attribute components whose similarities are greater than a component similarity threshold; and first transform coefficients of remaining attribute components of different to-be-encoded point cloud points in the third hybrid subsequence are consecutive, the remaining attribute components being attribute components other than the associated attribute components in the b attribute components; and the fourth hybrid subsequence comprises the second transform coefficients corresponding to the b attribute components of the to-be-encoded point cloud points, second transform coefficients of associated attribute components of a same to-be-encoded point cloud point in the fourth hybrid subsequence are adjacent, and second transform coefficients of remaining attribute components of different to-be-encoded point cloud points in the fourth hybrid subsequence are consecutive.
  10. The method according to claim 1, wherein the to-be-encoded point cloud points comprise at least two attribute parameters; the transform coefficients of the to-be-encoded point cloud points comprise first transform coefficients of the at least two attribute parameters, and second transform coefficients of the at least two attribute parameters; and in the to-be-encoded transform coefficient sequence, a first transform coefficient and a second transform coefficient parameter under a same attribute are consecutive, first transform coefficients under different attribute parameters are inconsecutive, and second transform coefficients under the different attribute parameters are inconsecutive.
  11. The method according to any one of claims 1 to 10, wherein the to-be-encoded transform coefficient sequence comprises F transform coefficients, the F transform coefficients comprising F1 first transform coefficients and F2 second transform coefficients, F being a positive integer; and F1 being a positive integer less than or equal to F, and F2 being a positive integer less than or equal to F; and the encoding the to-be-encoded transform coefficient sequence to obtain a point cloud group bitstream corresponding to the to-be-encoded point cloud group comprises: in a case that an ith transform coefficient in the F transform coefficients is a second transform coefficient, performing encoding processing on the ith transform coefficient to obtain an encoded codeword of the ith transform coefficient, i being a positive integer less than or equal to F; in a case that the ith transform coefficient is a first transform coefficient, determining a first transform coefficient in the F transform coefficients as a first coefficient predicted value, determining a first coefficient residual of the ith transform coefficient based on the first coefficient predicted value, and performing encoding processing on the first coefficient residual to obtain the encoded codeword of the ith transform coefficient; and in a case that i is F, using encoded codewords corresponding to the F transform coefficients to form the point cloud group bitstream corresponding to the to-be-encoded point cloud group.
  12. A point cloud encoding apparatus, characterized in that the apparatus comprises: an information obtaining module (34), configured to obtain encoding limit information; a sequence obtaining module (11), configured to obtain a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information; and a coefficient encoding module (12), configured to encode the to-be-encoded transform coefficient sequence to obtain a point cloud group bitstream corresponding to the to-be-encoded point cloud group, wherein the obtaining of a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information comprises: obtaining a group of to-be-encoded point cloud groups from point cloud groups based on the encoding limit information, wherein the encoding limit information comprises a group quantity limit threshold and a quantity of the to-be-encoded point cloud groups comprised in the group is less than or equal to the group quantity limit threshold, performing point cloud prediction processing on the to-be-encoded point cloud group to obtain attribute predicted values corresponding to to-be-encoded point cloud points included in the to-be-encoded point cloud group; obtaining attribute residuals of the to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values; obtaining transform coefficients of to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values and the obtained attribute residuals; and sorting transform coefficients of the to-be-encoded point cloud points included in the to-be-encoded point cloud group to obtain the to-be-encoded transform coefficient sequence.
  13. One or more computer-readable storage media, storing computer-readable instructions, the computer-readable instructions being suitable for loading and executing by a processor, to enable a computer device having the processor to perform the method according to any one of claims 1 to 11.
  14. One or more computer-readable storage media, storing computer-readable instructions, the computer-readable instructions being suitable for loading and executing by a first and a second processor, to enable a computer device having the first processor to perform the method according to any one of claims 1 to 11 and having the second processor to perform a point cloud decoding method characterized in that the point cloud decoding method comprises: obtaining the point cloud group bitstream (S501), the point cloud group bitstream being used for representing an encoded bitstream of a to-be-decoded point cloud group, a quantity of to-be-decoded point cloud groups being less than or equal to a group quantity limit threshold in encoding limit information; decoding the point cloud group bitstream, to obtain a transform coefficient corresponding to the to-be-decoded point cloud group (S502); wherein the encoding limit information comprising at least the group quantity limit threshold is obtained either by being signalled in encoded form within the point cloud group bitstream or by being pre-stored as part of predetermined parameters of the decoding device; and wherein the encoding limit information is used to determine a quantity of encoding bits and a context model for encoding or decoding a run length.

Description

RELATED APPLICATION This application claims priority to Chinese Patent Application No. 202210243506X, entitled "POINT CLOUD ENCODING AND DECODING METHOD AND APPARATUS, COMPUTER, AND STORAGE MEDIUM", filed with the China National Intellectual Property Administration on March 11, 2022. FIELD OF THE TECHNOLOGY The disclosure relates to the field of computer technologies, and in particular, to a point cloud encoding and decoding method and apparatus, a computer, and a storage medium. BACKGROUND OF THE DISCLOSURE In an existing point cloud attribute transform and predicting transform methods, all transform coefficients generated by point clouds are uniformly recorded in a very large memory, and then uniformly encoded, which increases memory overheads of encoding and decoding, and may further make an implementation of the process uncontrollable. In addition, uniform encoding of the transform coefficients leads to low efficiency in spatial random access of the point clouds and decoding and reconstruction. Patent literature CN 113 489 980 A (UNIV PEKING SHENZHEN GRADUATE SCHOOL) 8 October 2021 (2021-10-08) discloses methods of and devices for point cloud processing, in particular a method and an apparatus for entropy coding and entropy decoding of point cloud attribute transformation coefficients. Transform coefficients are processed in units of groups and group-wise entropy encoding is performed. SUMMARY Embodiments of the disclosure provide a point cloud encoding method and apparatus, a computer, and a storage medium. According to an aspect of the embodiments of the disclosure, a point cloud encoding method is provided, and is performed by a computer device. The method includes: obtaining encoding limit information, and obtaining a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information (S301) ; andencoding the to-be-encoded transform coefficient sequence to obtain a point cloud group bitstream corresponding to the to-be-encoded point cloud group (S303),wherein the obtaining of a to-be-encoded transform coefficient sequence of to-be-encoded point cloud points comprised in a to-be-encoded point cloud group based on the encoding limit information comprises: obtaining a group of to-be-encoded point cloud groups from point cloud groups based on the encoding limit information, wherein the encoding limit information comprises a group quantity limit threshold and a quantity of the to-be-encoded point cloud groups comprised in the group is less than or equal to the group quantity limit threshold,performing point cloud prediction processing on the to-be-encoded point cloud group to obtain attribute predicted values corresponding to to-be-encoded point cloud points included in the to-be-encoded point cloud group;obtaining attribute residuals of the to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values;obtaining transform coefficients of to-be-encoded point cloud points included in the to-be-encoded point cloud group based on the obtained attribute predicted values and the obtained attribute residuals; and sorting transform coefficients of the to-be-encoded point cloud points included in the to-be-encoded point cloud group to obtain the to-be-encoded transform coefficient sequence. According to an aspect of the embodiments of the disclosure, a point cloud encoding apparatus in accordance with the above-mentioned point cloud encoding method is provided. According to an aspect of the embodiments of the disclosure, a computer program product is provided, the computer program product including computer-readable instructions, the computer-readable instructions being stored in one or more computer-readable storage media. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and one or more processors execute the computer instructions, so that the computer device performs the method provided in the various optional implementations in an aspect of the embodiments of the disclosure. In other words, the computer-readable instructions, when executed by one or more processors, implement the method provided in various optional implementations in an aspect of the embodiments of the disclosure. Details of one or more embodiments of the disclosure are provided in the accompany drawings and descriptions below. Other features, objectives, and advantages of the disclosure become obvious with reference to the specification, the accompanying drawings, and the claims. BRIEF DESCRIPTION OF THE DRAWINGS To describe the technical solutions of the embodiments of the disclosure or the related art more clearly, the following briefly introduces the accompanying drawings required for describing the embodiments or the related art. Apparently, the accompanying drawings in the following description show only some embod