Granted Patent B2

US 12,423,877 · App. 18/287,839 · Granted Sep 23, 2025

Inventors: Sooyeon Lee (Seoul, KR); Hyunmook Oh (Seoul, KR); Hyejung Hur (Seoul, KR); Yousun Park (Seoul, KR)

Assignee: LG Electronics Inc.

G06T9/001

Abstract

Claims (64)

1. A method of transmitting point cloud data, the method comprising:

encoding geometry data containing positions of points of the point cloud data;

encoding attribute data containing attribute values of the points of the point cloud data based on the geometry data; and

transmitting the encoded geometry data, the encoded attribute data, and signaling information,

wherein the encoding of the geometry data comprises compressing the geometry data by performing motion estimation based on points selected in a search window of a reference frame and points selected in a prediction unit to be encoded of a current frame,

wherein the signaling information includes geometry compression related information, and

wherein the geometry compression related information includes at least motion vector information or information for point selection.

2. The method of claim 1 , wherein the encoding of the geometry data comprises:

acquiring first reference position information based on points included in the prediction unit to be encoded of the current frame;

selecting, based on the first reference position information, first points for motion estimation from among points included in the prediction unit;

acquiring second reference position information based on points included in the search window of the reference frame;

selecting, based on the second reference position information, second points for motion estimation from among the points included in the search window;

estimating a motion vector based on the first points and the second points; and

performing motion compensation based on the motion vector.

3. The method of claim 2 ,

wherein the first reference position information is a mean of the points included in the prediction unit, and

wherein the second reference position information is a mean of the points included in the search window.

4. The method of claim 2 ,

wherein the first reference position information is a median of the points included in the prediction unit, and

wherein the second reference position information is a median of the points included in the search window.

5. The method of claim 2 ,

wherein the selecting of the first points selects, from among the points included in the prediction unit, points within a first range from the first reference position information as the first points,

wherein the selecting of the second points selects, from among the points included in the search window, points within a second range from the second reference position information as the second points.

6. The method of claim 2 , wherein the estimating of the motion vector comprises:

acquiring first representative position information based on the first points;

acquiring second representative position information based on the second points; and

estimating a difference between the first representative position information and the second representative position information as the motion vector.

7. A device for transmitting point cloud data, the device comprising:

a geometry encoder configured to encode geometry data containing positions of points of point cloud data;

an attribute encoder configured to encode attribute data containing attribute values of the points of the point cloud data based on the geometry data; and

a transmitter configured to transmit the encoded geometry data, the encoded attribute data, and signaling information,

wherein the geometry encoder compresses the geometry data by performing motion estimation based on points selected in a search window of a reference frame and points selected in a prediction unit to be encoded of a current frame,

wherein the signaling information includes geometry compression related information, and

wherein the geometry compression related information includes at least motion vector information or information for point selection.

8. The device of claim 7 , wherein the geometry encoder comprises:

a first reference position information acquirer configured to acquire first reference position information based on points included in the prediction unit to be encoded of the current frame;

a first selector configured to select, based on the first reference position information, first points for motion estimation from among points included in the prediction unit;

a second reference position information acquirer configured to acquire second reference position information based on points included in a search window of the reference frame;

a second selector configured to select, based on the second reference position information, second points for motion estimation from among the points included in the search window;

a motion estimator configured to estimate a motion vector based on the first points and the second points; and

a motion compensator configured to perform motion compensation based on the motion vector.

9. The device of claim 8 ,

wherein the first reference position information is a mean of the points included in the prediction unit, and

wherein the second reference position information is a mean of the points included in the search window.

10. The device of claim 8 ,

wherein the first reference position information is a median of the points included in the prediction unit, and

wherein the second reference position information is a median of the points included in the search window.

11. The device of claim 8 ,

wherein the first selector selects, from among the points included in the prediction unit, points within a first range from the first reference position information as the first points,

wherein the second selector selects, from among the points included in the search window, points within a second range from the second reference position information as the second points.

12. The device of claim 8 , wherein the motion estimator is configured to:

acquire first representative position information based on the first points;

acquire second representative position information based on the second points; and

estimate a difference between the first representative position information and the second representative position information as the motion vector.

13. A method of receiving point cloud data, the method comprising:

receiving geometry data, and signaling information;

decoding the geometry data based on the signaling information; and

decoding the attribute data based on the signaling information and the decoded geometry data,

wherein the decoded geometry data includes positions of points for the point cloud data and the decoded attribute data includes attribute values of the points,

wherein the decoding of the geometry data comprises:

selecting points for motion compensation from among points included in a search window of a reference frame based on the signaling information, and

performing motion compensation based on points included in a prediction unit to be decoded of a current frame and the selected points in the search window,

wherein the signaling information includes geometry compression related information, and

wherein the geometry compression related information includes at least motion vector information or information for point selection.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (11)