Granted Patent B2

US 12,592,034 · App. 18/701,619 · Granted Mar 31, 2026

Inventors: Hong Shang (Beijing, CN); Xiang Li (Beijing, CN); Zhan Shi (Beijing, CN); Kuanhong Xu (Beijing, CN)

Assignee: SONY GROUP CORPORATION

G06T17/00G06T5/50G06T7/60G06T7/80G06T9/00G06V10/44G06V10/771G06T2207/20221

Abstract

Claims (35)

1 . A method for three-dimensional reconstruction, comprising:

obtaining a composite image obtained by photographing a body of a target object and a mirror image of the target object; and

performing three-dimensional reconstruction on the target object by using the composite image and relative position and posture information representing a position and posture of a mirror for generating the mirror image relative to a camera for the photography, wherein the performing three-dimensional reconstruction comprises

performing random sampling on points within a given three-dimensional space comprising the target object;

extracting a global feature corresponding to a sampling point from the composite image;

performing encoding on imaging-related geometric information of the sampling point, to generate geometric encoded information; and

inputting the global feature and the geometric encoded information into a model, to determine a geometric relationship between the sampling point and a surface of the target object.

2 . The method according to claim 1 , wherein the extracting the global feature corresponding to the sampling point from the composite image comprises:

performing feature extraction on the composite image, to obtain a global feature map;

determining a geometric association between the sampling point and a pixel point of the composite image; and

determining the global feature corresponding to the sampling point from the global feature map based on the geometric association.

3 . The method according to claim 2 , wherein the determining the geometric association between the sampling point and the pixel point of the composite image comprises:

determining a geometric association between the sampling point and a body pixel point in the composite image, which is obtained by photographing the body of the target object; and

determining, according to the relative position and posture information, a geometric association between the sampling point and a mirror image pixel point in the composite image, which is obtained by photographing the mirror image of the target object.

4 . The method according to claim 1 , further comprising:

calibrating at least one of the camera or the mirror, to obtain the relative position and posture information.

5 . A non-transitory computer-readable storage medium having one or more instructions stored thereon, which, when executed by a processor, cause the processor to perform the method for three-dimensional reconstruction according to claim 1 .

6 . A system for three-dimensional reconstruction, comprising:

an information processing apparatus comprising:

a memory; and

a processor coupled to the memory, the processor being configured to perform, based on instructions stored in the memory, the steps of the method according to claim 1 .

7 . The system according to claim 6 , further comprising:

a plane mirror group providing the mirror.

8 . The system according to claim 7 , wherein the plane mirror group is formed by plane mirror units.

9 . The system according to claim 7 , wherein at least one of the number of the plane mirror groups, or a position or posture of at least part of the plane mirror groups relative to the camera is adjustable as needed.

10 . The system according to claim 7 , wherein the plane mirror group comprises a position and posture obtaining module disposed thereon and configured to obtain and send information related to a position and posture of the plane mirror group.

11 . An electronic device, comprising:

a memory; and

a processor coupled to the memory, the processor being configured to perform, based on instructions stored in the memory, the steps of:

obtaining a composite image obtained by photographing a body of a target object and a mirror image of the target object; and

performing three-dimensional reconstruction on the target object by using the composite image and relative position and posture information representing a position and posture of a mirror for generating the mirror image relative to a camera for the photography, wherein the performing three-dimensional reconstruction comprises

performing random sampling on points within a given three-dimensional space comprising the target object;

extracting a global feature corresponding to a sampling point from the composite image;

performing encoding on imaging-related geometric information of the sampling point, to generate geometric encoded information; and

inputting the global feature and the geometric encoded information into a model, to determine a geometric relationship between the sampling point and a surface of the target object.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (15)