# US 10,636,206 B2 · Method and system for generating an image file of a 3D garment model on a 3D body model

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

- Number: US 10,636,206 B2
- Document ID: US10636206B2
- Status: Granted
- Kind: B2
- Application number: 15/898,365
- Filed: 2018-02-16
- Published: 2018-07-12
- Granted: 2020-04-28
- Earliest claimed: 2015-08-14
- Assignees: Metail Limited
- Applicants: Metail Limited
- Inventors: Yu Chen; Dongjoe Shin; Joe Townsend; Jim Downing; Duncan Robertson; Tom Adeyoola
- Web page: https://patentgrep.com/patent/US10636206B2

## CPC

- G06T 17/20: Finite element generation, e.g. wire-frame surface description, {tesselation}
- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06T 15/04: Texture mapping
- G06T 19/20: Editing of three-dimensional [3D] images, e.g. changing shapes or colours, aligning objects or positioning parts
- G06F 2113/12: Cloth
- G06T 2215/16: Using real world measurements to influence rendering
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method of generating an image file of a high resolution 3D garment model on a 3D body model, comprising the steps of: (i) acquiring at least three 2D images of a garment on a mannequin, wherein the at least three 2D images capture a 360° azimuthal angle range of view of the garment, including a texture of the garment; (ii) creating a simplified 3D garment model using the 2D images of the garment; (iii) simulating the smooth and simplified 3D garment model on the 3D body model; (iv) deforming and rendering the high-resolution garment texture on the simulated smooth and simplified 3D garment model on the 3D body model, and (v) generating an image file of the 3D body model, the image file including the deformed and rendered high-resolution garment texture, on the simulated smooth and simplified 3D garment model, on the 3D body model.

## Claims

### Claim 1

Method of using efficient caching and approximation strategies to reduce execution of explicit garment simulations at run-time, comprising the steps of:
(i) first pre-computing a deformation of a garment on a finite number of body shapes that are sampled from the human population, the pre-computing comprising:
(a) pre-computing explicit physics simulations of the garment on a finite number of body shapes;
(b) storing the 2D texture morph fields of each garment layer and camera view, based on the physics simulations; and
(c) interpolating the garment morph for the specified body shape from examples, at rendering time, wherein a piecewise bilinear interpolation is applied to generate approximate 2D garment morph fields from at least four nearby pre-computed texture morph samples, at rendering time, and storing these simulation results in a database;
(ii) for a specified body shape, generating an approximate garment deformation on the specified body shape by interpolating or extrapolating the pre-computed garment simulation results stored in the database on nearby sample body shapes to the specified body shape.

### Claim 2 (depends on 1)

Method of claim 1, wherein the stored 2D texture morph fields of each garment layer and camera view, are represented as rectangular grids, based on the physics simulations.

### Claim 3 (depends on 1)

Method of claim 1, wherein the interpolation of the garment morph for the specified body shape from examples is performed in a weighted-average manner.

### Claim 4 (depends on 1)

Method of claim 1, in which the physics simulations are 3D physics simulations.

### Claim 5 (depends on 1)

Method of claim 1, wherein interpolated texture morph fields are used directly to warp garment textures for a final rendering and garment texture composition for outfitting.

### Claim 6 (depends on 1)

Method of claim 1, wherein a multi-dimensional piecewise linear interpolation defined on K-different measurement dimensions is applied to generate approximate 2D garment morph fields.

### Claim 7 (depends on 1)

Method of claim 1, further comprising a method of generating an image file of a high resolution 3D garment model on a 3D body model, the method comprising the steps of:
(i) acquiring at least three 2D images of a garment on a mannequin, wherein the at least three 2D images of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing a texture of the garment at high resolution;
(ii) creating a smooth and simplified 3D garment model using the at least three 2D images of the garment;
(iii) simulating the smooth and simplified 3D garment model on the 3D body model;
(iv) deforming and rendering the high-resolution garment texture on the simulated smooth and simplified 3D garment model on the 3D body model, and
(v) generating an image file of the 3D body model, the image file including the deformed and rendered high-resolution garment texture, on the simulated smooth and simplified 3D garment model, on the 3D body model.

### Claim 8 (depends on 7)

Method of claim 7, in which the method includes: acquiring at least three depth scans of the garment on the mannequin, wherein the at least three depth scans of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing the texture of the garment at high resolution, and creating the smooth and simplified 3D garment model using the at least three depth scans of the garment.

### Claim 9 (depends on 1)

Method of claim 1, further comprising a method of generating an image file of a high resolution 3D garment model on a 3D body model, the method comprising the steps of:
(i) acquiring at least three 2D images of a garment on a mannequin, wherein the at least three 2D images of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing a texture of the garment at high resolution;
(ii) creating a smooth and simplified 3D garment model using the at least three 2D images of the garment;
(iii) simulating the smooth and simplified 3D garment model on the 3D body model;
(iv) deforming and rendering the high-resolution garment texture on the simulated smooth and simplified 3D garment model on the 3D body model, and
(v) generating an image file of the 3D body model, the image file including the deformed and rendered high-resolution garment texture, on the simulated smooth and simplified 3D garment model, on the 3D body model.

### Claim 10 (depends on 9)

Method of claim 9, in which the method includes: acquiring at least three depth scans of the garment on the mannequin, wherein the at least three depth scans of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing the texture of the garment at high resolution, and creating the smooth and simplified 3D garment model using the at least three depth scans of the garment.

### Claim 11 (depends on 1)

Method of claim 1, further comprising a method of creating a high quality texture map of a garment by stitching together multiple high definition (HD) images of the garment from a depth sensor, wherein the garment is on a mannequin and the mannequin is rotated on a turntable, under the control of a control unit, while the HD images are obtained from the depth sensor, which is under the control of the control unit.

### Claim 12 (depends on 1)

Method of claim 1, further comprising a method of creating a high quality texture map of a garment by stitching together multiple high definition (HD) images of the garment from a depth sensor, wherein the garment is on a mannequin and the mannequin is rotated on a turntable, under the control of a control unit, while the HD images are obtained from the depth sensor, which is under the control of the control unit.

### Claim 13 (depends on 1)

Method of claim 1, further comprising a method of garment texture morphing based on 2D physics simulation, in which a silhouette of a garment is deformed on a collision obstacle formed by a projection of a body model in a corresponding camera view.

### Claim 14 (depends on 1)

Method of claim 1, further comprising a method of garment texture morphing based on 2D physics simulation, in which a silhouette of a garment is deformed on a collision obstacle formed by a projection of a body model in a corresponding camera view.

### Claim 15

Computer program product embodied on a non-transitory storage medium, the computer program product executable on a processor to use efficient caching and approximation strategies to reduce execution of explicit garment simulations at run-time, the computer program product executable on the processor to:
(i) first pre-compute a deformation of a garment on a finite number of body shapes that are sampled from the human population, the pre-computing comprising:
(a) pre-computing explicit physics simulations of the garment on a finite number of body shapes;
(b) storing the 2D texture morph fields of each garment layer and camera view, based on the physics simulations; and
(c) interpolating the garment morph for the specified body shape from examples, at rendering time, wherein a piecewise bilinear interpolation is applied to generate approximate 2D garment morph fields from at least four nearby pre-computed texture morph samples, at rendering time, and storing these simulation results in a database;
(ii) for a specified body shape, generate an approximate garment deformation on the specified body shape by interpolating or extrapolating the pre-computed garment simulation results stored in the database on nearby sample body shapes to the specified body shape.

### Claim 16

Computer system including a processor, the computer system configured to use efficient caching and approximation strategies to reduce execution of explicit garment simulations at run-time, the computer system configured to:
(i) first pre-compute a deformation of a garment on a finite number of body shapes that are sampled from the human population, the pre-computing comprising:
(a) pre-computing explicit physics simulations of the garment on a finite number of body shapes;
(b) storing the 2D texture morph fields of each garment layer and camera view, based on the physics simulations; and
(c) interpolating the garment morph for the specified body shape from examples, at rendering time, wherein a piecewise bilinear interpolation is applied to generate approximate 2D garment morph fields from at least four nearby pre-computed texture morph samples, at rendering time, and storing these simulation results in a database;
(ii) for a specified body shape, generate an approximate garment deformation on the specified body shape by interpolating or extrapolating the pre-computed garment simulation results stored in the database on nearby sample body shapes to the specified body shape.

### Claim 17

Method of using efficient caching and approximation strategies to reduce execution of explicit garment simulations at run-time, and generating an image file of a high resolution 3D garment model on a 3D body, comprising the steps of:
(i) acquiring at least three 2D images of a garment on a mannequin, wherein the at least three 2D images of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing a texture of the garment at high resolution;
(ii) creating a smooth and simplified 3D garment model using the at least three 2D images of the garment;
(iii) simulating the smooth and simplified 3D garment model on the 3D body model;
(iv) deforming and rendering the high-resolution garment texture on the simulated smooth and simplified 3D garment model on the 3D body model; and
(v) generating an image file of the 3D body model, the image file including the deformed and rendered high-resolution garment texture, on the simulated smooth and simplified 3D garment model, on the 3D body model;
(vi) pre-computing a deformation of a garment on a finite number of body shapes that are sampled from the human population, and storing these simulation results in a database;
(vii) for a specified body shape, generating an approximate garment deformation on the specified body shape by interpolating or extrapolating the pre-computed garment simulation results stored in the database on nearby sample body shapes to the specified body shape.

### Claim 18 (depends on 17)

Method of claim 17, further comprising:
(a) pre-computing explicit physics simulations of the garment on a finite number of body shapes;
(b) storing the 2D texture morph fields of each garment layer and camera view, based on the physics simulations; and
(c) interpolating the garment morph for the specified body shape from examples, at rendering time.

### Claim 19 (depends on 17)

Method of claim 17, in which the method includes: acquiring at least three depth scans of the garment on the mannequin, wherein the at least three depth scans of the garment on the mannequin capture a 360° azimuthal angle range of view of the garment on the mannequin, including capturing the texture of the garment at high resolution, and creating the smooth and simplified 3D garment model using the at least three depth scans of the garment.
