# US 11,443,480 B2 · Method and system for remote clothing selection

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

- Number: US 11,443,480 B2
- Document ID: US11443480B2
- Status: Granted
- Kind: B2
- Application number: 17/081,912
- Filed: 2020-10-27
- Published: 2021-03-18
- Granted: 2022-09-13
- Earliest claimed: 2019-08-13
- Assignees: Texel Inc.
- Applicants: Texel LLC
- Inventors: Maxim Alexandrovich Fedyukov; Andrey Vladimirovich Poskonin; Sergey Mikhailovich Klimentyev; Vladimir Vladimirovich Guzov; Ilia Alexeevich Petrov; Nikolay Patakin; Anton Vladimirovich Fedotov; Oleg Vladimirovich Korneev
- Web page: https://patentgrep.com/patent/US11443480B2

## CPC

- G06T 17/00: Three-dimensional [3D] modelling for computer graphics
- G06Q 30/0643: {graphically representing goods, e.g. 3D product representation}
- G06T 7/50: Depth or shape recovery
- G06T 7/60: Analysis of geometric attributes
- G06T 17/20: Finite element generation, e.g. wire-frame surface description, {tesselation}
- G06T 19/20: Editing of three-dimensional [3D] images, e.g. changing shapes or colours, aligning objects or positioning parts
- G06T 2200/04: involving 3D image data
- G06T 2207/10028: Range image; Depth image; 3D point clouds
- G06T 2207/20028: Bilateral filtering
- G06T 2207/30196: Human being; Person
- G06T 2210/56: Particle system, point based geometry or rendering
- G06T 2219/2021: Shape modification
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

Method for remote clothing selection includes determination of anthropometric dimensional parameters of a user, automatic assessment of correspondence of a garment to the shape and body measurements of a user, determination and provision of recommendations to a user on the selection of a particular garment and, optionally, visualization of a garment on a digital avatar of this user in the virtual fitting room, including optional change of his/her pose. The invention provides an increase in the efficiency of remote clothing selection by a user, an improvement in user's experience of remote purchase, an increase in user satisfaction and, ultimately, an increase in online sales of clothing and a decrease in the proportion of clothing returned after a purchase due to unsatisfactory matching to the shape and measurements of user's body.

## Claims

### Claim 1

A method for determining dimensional parameters of a garment, the method comprising the steps of:
obtaining a geometrical production model used for manufacturing of the garment;
transforming the geometrical production model into a parametric production model based on a Non-Rigid Deformation algorithm comprising a combined loss function as follows:
E=αEdata+γEsmooth+βEnormal,
where Edata is a data loss function, Esmooth is a smoothness loss function, Enormal is a loss function which penalizes large angles between normals to neighboring triangles, and α, β, γ are weight factors; and
deriving the dimensional parameters of the garment from the parametric production model.

### Claim 2 (depends on 1)

The method of claim 1, wherein the geometrical production model is obtained based on a tailor's dummy.

### Claim 3 (depends on 1)

The method of claim 1, wherein the geometrical production model is obtained based on a human model.

### Claim 4 (depends on 1)

The method of claim 1, wherein the geometrical production model is obtained based on a computer model.

### Claim 5 (depends on 1)

The method of claim 1, wherein the geometrical production model is obtained based on at least one of a spec sheet, a tech pack, and a set of grading rules.

### Claim 6 (depends on 1)

The method of claim 1, wherein the Edata loss function is
E
data
=
∑
i
=
1
N
⁢
w
i
⁢

A
i
⁢
p
i
-
N
⁢
N
⁡
(
i
)

2
,
where wi is a weight factor of an ith vertex in the parametric production model, Ai is a transformation matrix for the ith vertex, pi is a coordinate of the ith vertex in the parametric production model, NN(i) is a vertex of the geometrical production model being a neighbor of the ith vertex in the parametric production model.

### Claim 7 (depends on 1)

The method of claim 1, wherein the Esmooth loss function is
E
smooth
=
∑
(
i
,
j
)
∈
edges
⁡
(
T
)
⁢

A
i
-
A
j

F
2
,
where edges(T) is a set of edges in the parametric production model, Ai is a 4×4 transformation matrix for an ith vertex, Ai is a 4×4 transformation matrix for a jth vertex, ∥⋅∥F is a Frobenius norm of a corresponding matrix.

### Claim 8 (depends on 1)

The method of claim 1, wherein the Enormal loss function is
E
normal
=
1

edges

⁢
∑
(
i
,
j
)
∈
edges
⁡
(
T
)
⁢
⁢
F
⁡
(
)
,
where |edges| is a number of edges in the parametric production model, edges(T) is a set of edges in the parametric production model, ni is a normal vector to an ith triangle, nj is a vector of normal to a jth triangle, is an angle between normal vectors to ith and jth triangles, and F is an activation function:
F
⁡
(
α
)
=
{
x
3
90
3
0
⁢
°
≤
α
≤
90
⁢
°
0.5
+
α
180
α
≥
90
⁢
°
.

### Claim 9 (depends on 1)

The method of claim 1, wherein the weight factor α is equal to a first predetermined value at the beginning of the algorithm and then is changed to a second predetermined value.

### Claim 10 (depends on 9)

The method of claim 9, wherein the first predetermined value is 0 and the second predetermined value is 1.

### Claim 11 (depends on 1)

The method of claim 1, wherein the weight factor β is equal to a first predetermined value initially and then is changed to a second predetermined value.

### Claim 12 (depends on 11)

The method of claim 11, wherein the first predetermined value is equal to 10−3 and the second predetermined value is reduced by four times of its previous value at each step.

### Claim 13 (depends on 12)

The method of claim 12, wherein the second predetermined value is adjusted until β≥10−6α or until the algorithm makes no further progress.

### Claim 14 (depends on 1)

The method of claim 1, wherein the weight factor γ is equal to a first predetermined value initially and then is changed to a second predetermined value.

### Claim 15 (depends on 14)

The method of claim 14, wherein the first predetermined value is equal to 106 and the second predetermined value is reduced by four times of its previous value at each step.

### Claim 16 (depends on 15)

The method of claim 15, wherein the second predetermined value is adjusted until β≥10−6α or until the algorithm makes no further progress.