# US 2026/0073653 A1 · Visual guided mesh repair

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0073653 A1
- Document ID: US20260073653A1
- Status: Published application
- Kind: A1
- Application number: 19/394,616
- Filed: 2025-11-19
- Published: 2026-03-12
- Earliest claimed: 2023-07-26
- Assignees: Tencent America LLC
- Applicants: Tencent America LLC
- Inventors: Kui WU; Xifeng GAO; Zherong PAN
- Web page: https://patentgrep.com/patent/US20260073653A1

## CPC

- G06T 19/20: Editing of three-dimensional [3D] images, e.g. changing shapes or colours, aligning objects or positioning parts
- G06T 15/06: Ray-tracing
- G06T 17/205: {Re-meshing}
- G06T 2210/12: Bounding box
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Aspects of the disclosure provide methods and apparatuses for mesh processing. In some examples, a method for mesh processing includes receiving an input mesh that is bounded in a bounding box. The input mesh is formed of a first plurality of vertices and a first plurality of faces that connects the first plurality of vertices. The method further includes calculating at least a first visual measure for one or more faces in the first plurality of faces. At least the first visual measure for a face is calculated according to a number of valid rays respectively at a plurality of sampling positions of the face, a valid ray at a sampling position is a ray that is traced from the sampling position to the bounding box. The method also includes performing adjustments to the input mesh to generate a repaired mesh according to at least the first visual measure.

## Claims

### Claim 1

A method of mesh processing, comprising:
receiving an input mesh that is bounded in a bounding box, the input mesh being formed of a first plurality of vertices and a first plurality of faces that connects the first plurality of vertices;
calculating at least one first visual measure for one or more faces in the first plurality of faces, the at least one first visual measure for a face in the one or more faces being calculated according to a number of valid rays respectively at a plurality of sampling positions of the face, a valid ray at a sampling position of the face being a ray that is traced, within the bounding box of the input mesh, from the sampling position of the face to the bounding box of the input mesh; and
performing adjustments to the input mesh to generate a repaired mesh according to the at least one first visual measure.

### Claim 2 (depends on 1)

The method of claim 1, wherein the calculating further comprises:
determining a first number of valid rays at a first sampling position at a positive side of the face, the positive side being pointed by a face normal of the face;
determining a second number of valid rays at the first sampling position at a negative side of the face, the negative side being opposite of the positive side; and
calculating the at least one first visual measure of the face according to at least the first number of valid rays and the second number of valid rays.

### Claim 3 (depends on 2)

The method of claim 2, wherein the at least one first visual measure comprises a visibility measure, and the calculating the at least one first visual measure comprises:
determining, among first numbers of valid rays respectively associated with the plurality of sampling positions at the positive side of the face and second numbers of valid rays respectively associated with the plurality of sampling positions at the negative side of the face, a maximum number of valid rays; and
calculating the visibility measure for the face based on the maximum number of valid rays.

### Claim 4 (depends on 2)

The method of claim 2, wherein the at least one first visual measure comprises an orientation measure, and the calculating the at least one first visual measure comprises:
calculating a difference of a first total number of valid rays at the positive side of the face and a second total number of valid rays at the negative side of the face; and
calculating the orientation measure for the face based on the difference.

### Claim 5 (depends on 2)

The method of claim 2, wherein the at least one first visual measure comprises an openness measure, and the calculating the at least one first visual measure comprises:
calculating a ratio between a smaller one of the first number of valid rays and the second number of valid rays and a larger one of the first number of valid rays and the second number of valid rays; and
calculating a value of the openness measure based on the ratio.

### Claim 6 (depends on 4)

The method of claim 4, wherein the performing the adjustments to the input mesh further comprises:
determining that the orientation measure for the face is indicative of the face being misoriented; and
when the face is misoriented, flipping the face normal of the face to an opposite direction for an orientation adjustment of the face.

### Claim 7 (depends on 5)

The method of claim 5, wherein the performing the adjustments to the input mesh further comprises:
determining that the openness measure for the face indicates the face being an open surface; and
when the face is the open surface, forming a closed shell for an openness adjustment of the face, the closed shell being formed by vertices of the face and additional vertices that offset the vertices of the face by a distance.

### Claim 8 (depends on 7)

The method of claim 7, wherein the additional vertices are offset from the vertices of the face in an opposite direction of the face normal by the distance.

### Claim 9 (depends on 3)

The method of claim 3, wherein the performing the adjustments to the input mesh further comprises:
generating an intermediate mesh with initial adjustments to the input mesh, the intermediate mesh being formed by a second plurality of vertices and a second plurality of faces that connects the second plurality of vertices, the second plurality of faces including the first plurality of faces;
performing an initial space partition of a volume in the bounding box based on the second plurality of vertices;
performing an additional space partition of the volume in the bounding box according to the second plurality of faces, the volume being partitioned into cells by a plurality of splitting faces, the plurality of splitting faces including the second plurality of faces;
classifying the plurality of splitting faces into visible faces, invisible faces and extra faces based on the visibility measure for the plurality of splitting faces; and
determining an interface mesh from the plurality of splitting faces, the interface mesh maximizing a use of the visible faces and minimizing a use of the extra faces.

### Claim 10 (depends on 9)

The method of claim 9, wherein the determining the interface mesh further comprises:
determining classifications for the cells that minimize a cost function, each of the cells being classified into one of an interior cell and an exterior cell, the cost function having a first portion that penalizes incorrectly oriented faces in the visible faces based on the classifications, and a second portion that penalizes a use of one or more faces from the extra faces in the interface mesh that borders interior cells with exterior cells.

### Claim 11 (depends on 1)

The method of claim 1, wherein the calculating the at least one first visual measure further comprises:
determining the plurality of sampling positions of the face based on an area of the face and a minimum sampling number for each face.

### Claim 12

An apparatus, comprising processing circuitry configured to:
receive an input mesh that is bounded in a bounding box, the input mesh being formed of a first plurality of vertices and a first plurality of faces that connects the first plurality of vertices;
calculate at least one first visual measure for one or more faces in the first plurality of faces, the at least one first visual measure for a face in the one or more faces being calculated according to a number of valid rays respectively at a plurality of sampling positions of the face, a valid ray at a sampling position of the face being a ray that is traced, within the bounding box of the input mesh, from the sampling position of the face to the bounding box of the input mesh; and
perform adjustments to the input mesh to generate a repaired mesh according to the at least one first visual measure.

### Claim 13 (depends on 12)

The apparatus of claim 12, wherein the processing circuitry is configured to:
determine a first number of valid rays at a first sampling position at a positive side of the face, the positive side being pointed by a face normal of the face;
determine a second number of valid rays at the first sampling position at a negative side of the face, the negative side being opposite of the positive side; and
calculate the at least one first visual measure of the face according to at least the first number of valid rays and the second number of valid rays.

### Claim 14 (depends on 13)

The apparatus of claim 13, wherein the at least one first visual measure comprises a visibility measure, and the processing circuitry is configured to:
determine, among first numbers of valid rays respectively associated with the plurality of sampling positions at the positive side of the face and second numbers of valid rays respectively associated with the plurality of sampling positions at the negative side of the face, a maximum number of valid rays; and
calculate the visibility measure for the face based on the maximum number of valid rays.

### Claim 15 (depends on 13)

The apparatus of claim 13, wherein the at least one first visual measure comprises an orientation measure, and the processing circuitry is configured to:
calculate a difference of a first total number of valid rays at the positive side of the face and a second total number of valid rays at the negative side of the face; and
calculate the orientation measure for the face based on the difference.

### Claim 16 (depends on 13)

The apparatus of claim 13, wherein the at least one first visual measure comprises an openness measure, and the processing circuitry is configured to:
calculate a ratio between a smaller one of the first number of valid rays and the second number of valid rays and a larger one of the first number of valid rays and the second number of valid rays; and
calculate a value of the openness measure based on the ratio.

### Claim 17 (depends on 15)

The apparatus of claim 15, wherein the processing circuitry is configured to:
determine that the orientation measure for the face is indicative of the face being misoriented; and
when the face is misoriented, flip the face normal of the face to an opposite direction for an orientation adjustment of the face.

### Claim 18 (depends on 16)

The apparatus of claim 16, wherein the processing circuitry is configured to:
determine that the openness measure for the face indicates the face being an open surface; and
when the face is the open surface, form a closed shell for an openness adjustment of the face, the closed shell being formed by vertices of the face and additional vertices that offset the vertices of the face by a distance.

### Claim 19 (depends on 14)

The apparatus of claim 14, wherein the processing circuitry is configured to:
generate an intermediate mesh with initial adjustments to the input mesh, the intermediate mesh being formed by a second plurality of vertices and a second plurality of faces that connects the second plurality of vertices, the second plurality of faces including the first plurality of faces;
perform an initial space partition of a volume in the bounding box based on the second plurality of vertices;
perform an additional space partition of the volume in the bounding box according to the second plurality of faces, the volume being partitioned into cells by a plurality of splitting faces, the plurality of splitting faces including the second plurality of faces;
classify the plurality of splitting faces into visible faces, invisible faces and extra faces based on the visibility measure for the plurality of splitting faces; and
determine an interface mesh from the plurality of splitting faces, the interface mesh maximizing a use of the visible faces and minimizing a use of the extra faces.

### Claim 20 (depends on 19)

The apparatus of claim 19, wherein the processing circuitry is configured to:
determine classifications for the cells that minimize a cost function, each of the cells being classified into one of an interior cell and an exterior cell, the cost function having a first portion that penalizes incorrectly oriented faces in the visible faces in response to the classifications, and a second portion that penalizes a use of one or more faces from the extra faces in the interface mesh that borders interior cells with exterior cells.
