IP Library › Granted Patent US 12,445,647
Granted Patent B2
US 12,445,647 · App. 18/265,850 · Granted Oct 14, 2025

Method and apparatus for encoding/decoding point cloud coordinate transformation residual

Inventors: Wei Zhang (Shenzhen, CN); Fuzheng Yang (Shenzhen, CN); Tian Chen (Shenzhen, CN); Wenjie Zou (Shenzhen, CN)
Assignee: HONOR DEVICE CO., LTD.
H04N19/60
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,445,647
App. No.
18/265,850
Filed
Jun 7, 2023
Granted
Oct 14, 2025
Kind
B2
Art Unit
2896
USPC
375/240.18
Abstract

An encoding method includes: obtaining original point cloud data; transforming coordinates of the original point cloud data to obtain a coordinate transformation residual; establishing a relationship between azimuth information of a point cloud and the coordinate transformation residual; determining an encoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and encoding the coordinate transformation residual according to the encoding rule, to obtain data rate information.

Claims (420)

1. A method for encoding a point cloud coordinate transformation residual, comprising:

obtaining original point cloud data;

transforming coordinates of the original point cloud data to obtain a coordinate transformation residual;

establishing a relationship between azimuth information of a point cloud and the coordinate transformation residual, wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

establishing function expressions between the azimuth information of the point cloud and the coordinate transformation residual as follows:

Δ

⁢

x

=

r

·

cos

⁢

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

cos

⁢

(

φ

+

Δφ

)

,

and

Δ

⁢

y

=

r

·

sin

⁢

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

sin

⁢

(

φ

+

Δφ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, r represents depth information of the point, φ represents azimuth information of the point, Δr represents a precision loss of the depth information of the point, and Δφ represents a precision loss of the azimuth information of the point;

determining an encoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

encoding the coordinate transformation residual according to the encoding rule, to obtain data rate information.

2. The method for encoding a point cloud coordinate transformation residual according to claim 1 , wherein the function expressions between the azimuth information of the point cloud and the coordinate transformation residual are approximated as follows:

Δ

⁢

x

≈

-

Δ

⁢

r

·

cos

⁢

(

φ

)

,

and

Δ

⁢

y

≈

-

Δ

⁢

r

·

sin

⁢

(

φ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, φ represents azimuth information of the point, and Δr represents a precision loss of depth information of the point.

3. The method for encoding a point cloud coordinate transformation residual according to claim 1 , wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual further comprises:

obtaining a specific representation form of an angle value of an azimuth, a function value of an azimuth, or an azimuth symbol based on a function expression of the azimuth information of the point cloud and the coordinate transformation residual.

4. The method for encoding a point cloud coordinate transformation residual according to claim 1 , wherein determining the encoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

dividing an azimuth into several intervals based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

setting a corresponding encoding rule for the coordinate transformation residual for each azimuth interval.

5. The method for encoding a point cloud coordinate transformation residual according to claim 4 , wherein setting the corresponding encoding rule for the coordinate transformation residual for each azimuth interval comprises:

separately assigning a context model for a coordinate transformation residual of each azimuth interval, and performing corresponding initialization for entropy encoding of the coordinate transformation residual of each azimuth interval.

6. The method for encoding a point cloud coordinate transformation residual according to claim 5 , wherein encoding the coordinate transformation residual according to the encoding rule comprises:

obtaining azimuth information of a current point, and determining an azimuth interval to which the current point belongs; and

for different coordinate transformation residual components of the current point, selecting a corresponding context model based on the azimuth interval to perform entropy encoding.

7. An apparatus for encoding a point cloud coordinate transformation residual, comprising:

a computing device, configured to:

obtain original point cloud data;

transform coordinates of the original point cloud data to obtain a coordinate transformation residual;

establish a relationship between azimuth information of a point cloud and the coordinate transformation residual, wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

establishing function expressions between the azimuth information of the point cloud and the coordinate transformation residual as follows:

Δ

⁢

x

=

r

·

cos

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

cos

⁡

(

φ

+

Δφ

)

,

and

Δ

⁢

y

=

r

·

sin

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

sin

⁡

(

φ

+

Δφ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, r represents depth information of the point, φ represents azimuth information of the point, Δr represents a precision loss of the depth information of the point, and Δφ represents a precision loss of the azimuth information of the point;

determine an encoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

encode the coordinate transformation residual according to the encoding rule, to obtain data rate information.

8. A method for decoding a point cloud coordinate transformation residual, comprising:

obtaining data rate information;

determining a decoding rule for the coordinate transformation residual based on a relationship between azimuth information of a point cloud and the coordinate transformation residual, wherein the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

function expressions between the azimuth information of the point cloud and the coordinate transformation residual are as follows:

Δ

⁢

x

=

r

·

cos

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

cos

⁡

(

φ

+

Δφ

)

,

and

Δ

⁢

y

=

r

·

sin

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

sin

⁡

(

φ

+

Δφ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, r represents depth information of the point, φ represents azimuth information of the point, Δr represents a precision loss of the depth information of the point, and Δφ represents a precision loss of the azimuth information of the point; and

decoding the data rate information according to the decoding rule, to obtain a reconstructed point cloud coordinate transformation residual.

9. An apparatus for decoding a point cloud coordinate transformation residual, comprising:

a computing device, configured to:

obtain data rate information;

determine a decoding rule for the coordinate transformation residual based on a relationship between azimuth information of a point cloud and the coordinate transformation residual, wherein the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

function expressions between the azimuth information of the point cloud and the coordinate transformation residual are as follows:

Δ

⁢

x

=

r

·

cos

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

cos

⁡

(

φ

+

Δφ

)

,

and

Δ

⁢

y

=

r

·

sin

⁡

(

φ

)

-

(

r

+

Δ

⁢

r

)

·

sin

⁡

(

φ

+

Δφ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, r represents depth information of the point, φ represents azimuth information of the point, Δr represents a precision loss of the depth information of the point, and Δφ represents a precision loss of the azimuth information of the point; and

decode the data rate information according to the decoding rule, to obtain a reconstructed point cloud coordinate transformation residual.

10. The apparatus for encoding a point cloud coordinate transformation residual according to claim 7 , wherein the function expressions between the azimuth information of the point cloud and the coordinate transformation residual are approximated as follows:

Δ

⁢

x

≈

-

Δ

⁢

r

·

cos

⁡

(

φ

)

,

and

Δ

⁢

y

≈

-

Δ

⁢

r

·

sin

⁡

(

φ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, φ represents azimuth information of the point, and Δr represents a precision loss of depth information of the point.

11. The apparatus for encoding a point cloud coordinate transformation residual according to claim 7 , wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual further comprises:

obtaining a specific representation form of an angle value of an azimuth, a function value of an azimuth, or an azimuth symbol based on a function expression of the azimuth information of the point cloud and the coordinate transformation residual.

12. The apparatus for encoding a point cloud coordinate transformation residual according to claim 7 , wherein determining the encoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

dividing an azimuth into several intervals based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

setting a corresponding encoding rule for the coordinate transformation residual for each azimuth interval.

13. The apparatus for encoding a point cloud coordinate transformation residual according to claim 12 , wherein setting the corresponding encoding rule for the coordinate transformation residual for each azimuth interval comprises:

separately assigning a context model for a coordinate transformation residual of each azimuth interval, and performing corresponding initialization for entropy encoding of the coordinate transformation residual of each azimuth interval.

14. The apparatus for encoding a point cloud coordinate transformation residual according to claim 13 , wherein encoding the coordinate transformation residual according to the encoding rule comprises:

obtaining azimuth information of a current point, and determining an azimuth interval to which the current point belongs; and

for different coordinate transformation residual components of the current point, selecting a corresponding context model based on the azimuth interval to perform entropy encoding.

15. The method for decoding a point cloud coordinate transformation residual according to claim 8 , wherein the function expressions between the azimuth information of the point cloud and the coordinate transformation residual are approximated as follows:

Δ

⁢

x

≈

-

Δ

⁢

r

·

cos

⁡

(

φ

)

,

and

Δ

⁢

y

≈

-

Δ

⁢

r

·

sin

⁡

(

φ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, φ represents azimuth information of the point, and Δr represents a precision loss of depth information of the point.

16. The apparatus for decoding a point cloud coordinate transformation residual according to claim 9 , wherein the function expressions between the azimuth information of the point cloud and the coordinate transformation residual are approximated as follows:

Δ

⁢

x

≈

-

Δ

⁢

r

·

cos

⁡

(

φ

)

,

and

Δ

⁢

y

≈

-

Δ

⁢

r

·

sin

⁡

(

φ

)

;

wherein Δx and Δy represent coordinate transformation residual components of a point, φ represents azimuth information of the point, and Δr represents a precision loss of depth information of the point.

17. The method for decoding a point cloud coordinate transformation residual, according to claim 8 , wherein determining the decoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

dividing an azimuth into several intervals based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

setting a corresponding decoding rule for the coordinate transformation residual for each azimuth interval.

18. The apparatus for decoding a point cloud coordinate transformation residual according to claim 9 , wherein determining the decoding rule for the coordinate transformation residual based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual comprises:

dividing an azimuth into several intervals based on the relationship between the azimuth information of the point cloud and the coordinate transformation residual; and

setting a corresponding decoding rule for the coordinate transformation residual for each azimuth interval.

19. The method for encoding a point cloud coordinate transformation residual according to claim 1 , wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual further comprises:

obtaining a specific representation form of an angle value of an azimuth based on a function expression of the azimuth information of the point cloud and the coordinate transformation residual.

20. The method for encoding a point cloud coordinate transformation residual according to claim 1 , wherein establishing the relationship between the azimuth information of the point cloud and the coordinate transformation residual further comprises:

obtaining a specific representation form of an azimuth symbol based on a function expression of the azimuth information of the point cloud and the coordinate transformation residual.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 1, 2023
From: ZHANG, WEI; YANG, FUZHENG; CHEN, TIAN; ZOU, WENJIE
To: HONOR DEVICE CO., LTD.
Reel/Frame 065737/0670 →
Priority Claims (1)
CN 202110580212.1 · May 26, 2021 · national
Continuity (1)
Related Publication 20250080772A1 · Mar 6, 2025
References Cited (24)
US 11398042B2 · Zhao · 2022 [cited by applicant]
US 20190311500A1 · Mammou et al. · 2019 [cited by applicant]
US 20200137399A1 · Li et al. · 2020 [cited by applicant]
US 20210049828A1 · Park et al. · 2021 [cited by applicant]
US 20220358686A1 · Lasserre et al. · 2022 [cited by applicant]
US 20220398784A1 · Taquet et al. · 2022 [cited by applicant]
US 20230196625A1 · Li et al. · 2023 [cited by applicant]
US 20230267651A1 · Zhang et al. · 2023 [cited by applicant]
US 20230328285A1 · Oh · 2023 [cited by examiner]
US 20240029311A1 · Park · 2024 [cited by examiner]
US 20240163451A1 · Gao · 2024 [cited by examiner]
CN 108876926A · 2018 [cited by applicant]
CN 111145090A · 2020 [cited by applicant]
CN 111699683A · 2020 [cited by applicant]
CN 112385236A · 2021 [cited by applicant]
CN 112565734A · 2021 [cited by applicant]
WO 2020248187A1 · 2020 [cited by applicant]
WO 2021022621A1 · 2021 [cited by applicant]
WO 2021084292A1 · 2021 [cited by applicant]
WO 2021084295A1 · 2021 [cited by applicant]
WG7, MPEG 3D Graphics Coding, “G-PCC codec description”, 133. MPEG Meeting, Jan. 11, 2021-Jan. 15, 2021, Online, (Motion Picture Expert Group or ISO/IEC JTC1/SC29/WG11), No. n20028, Apr. 22, 2021 (Apr. 22, 2021), 150 pa… [cited by applicant]
Jonathan Taquet (Xiaomi) et al: “[G-PCC][New] Improved Quantization of Azimuthal Angle in Predictive Geometry Coding”, 133. MPEG Meeting, Jan. 11, 2021-Jan. 15, 2021, Online, (Motion Picture Expert Group or ISO/IEC JTC1… [cited by applicant]
Geert Van Der Auwera (Qualcomm) et al: “[GPCC][New] Planar and azimuthal coding mode simplifications”, 131. MPEG Meeting, Jun. 29, 2020-Jul. 3, 2020, Online, (Motion Picture Expert Group or ISO/IEC JTC1/SC29/WG11), No. … [cited by applicant]
Sun, Xuebin et al., “A Novel Coding Architecture for LiDAR Point Cloud Sequence,” IEEE Robotics and Automation Letters, Oct. 2020, 8 pages, vol. 5, No. 4. [cited by applicant]