Wrinkle detection method and electronic device

Patent metadata

Field	Value
Publication number	US-11941804-B2
Application number	US-201817260015-A
Country	US
Kind code	B2
Filing date	Sep 18, 2018
Priority date	Jul 16, 2018
Publication date	Mar 26, 2024
Grant date	Mar 26, 2024

Abstract

A wrinkle detection method includes: rotating a region, in a face image, in which a wrinkle needs to be detected, to obtain a plurality of to-be-detected images; determining wrinkle points from all pixel points based on grayscale values of the pixel points in each of the to-be-detected images with different angles; determining at least one wrinkle line based on the wrinkle points; and then displaying, by the electronic device, the wrinkle line in the region in which the wrinkle needs to be detected, where each wrinkle line indicates one wrinkle in each of the to-be-detected images.

First claim

The invention claimed is:

1. A wrinkle detection method, comprising:
rotating, according to some or all preset angles in a preset angle set, a region of a face image, in which a wrinkle will be detected, to obtain a plurality of to-be-detected images, wherein the preset angle set comprises a plurality of preset angles, and values of the preset angles are different;
determining wrinkle points from all pixel points of each to-be-detected image of the plurality of to-be-detected images based on grayscale values of the pixel points in each to-be-detected image of the plurality of to-be-detected images;
determining at least one wrinkle line from the plurality of to-be-detected images based on the wrinkle points, wherein each of the at least one wrinkle line indicates one wrinkle in each to-be-detected image of the plurality of to-be-detected images;
performing, for the at least one wrinkle line corresponding to each to-be-detected image, reverse rotation according to a preset angle based on which the to-be-detected image is obtained, to obtain rotated wrinkle lines;
fusing the rotated wrinkle lines into the region in which the wrinkle will be detected, to obtain a wrinkle line corresponding to the region in which the wrinkle will be detected;
displaying the wrinkle line in the region;
determining a wrinkle score corresponding to the region in which the wrinkle will be detected based on a feature set of the displayed wrinkle line, wherein the feature set comprises the following features: a length of the wrinkle line, a width of the wrinkle line, and a contrast ratio value of a pixel point on the wrinkle line, and an area proportion of the wrinkle line, wherein the contrast ratio value represents a contrast ratio of the pixel point on the wrinkle line, and the area proportion of the wrinkle line represents a proportion of a quantity of pixels on the wrinkle line to a quantity of all pixels in the to-be-detected image; and
outputting the wrinkle score, wherein the wrinkle score is determined according to the following formula:
H=A×ω 1+ B×ω 2+ C×ω 3+ D×ω 4+ω5, wherein

H is the wrinkle score, A is the length of the wrinkle line, B is the width of the wrinkle line, C is the contrast ratio value of the pixel point on the wrinkle line, D is the area proportion of the wrinkle line, ω1, ω2, ω3, and ω4 are preset parameters less than o, and ω5 is a preset parameter.

2. The method of claim 1 , wherein determining the wrinkle points from all the pixel points of each to-be-detected image of the plurality of to-be-detected images based on the grayscale values of the pixel points in each to-be-detected image of the plurality of to-be-detected images comprises:
setting a rectangular window;
controlling, according to a specified sliding step, the rectangular window to traverse each to-be-detected image of the plurality of to-be-detected images;
determining, at each window position in which the rectangular window is located, a central pixel point located at a central position of the rectangular window, and determining a confidence value of the central pixel point based on grayscale values of all pixel points in the rectangular window, to obtain a plurality of confidence values, wherein each confidence value indicates a possibility that the central pixel point is the wrinkle point; and
using a central pixel point corresponding to a confidence value not less than a threshold as the wrinkle point.

3. The method of claim 2 , wherein both a length and a width of the rectangular window are N pixels; and determining the central pixel point located at the central position of the rectangular window, and determining the confidence value of the central pixel point based on the grayscale values of all pixel points in the rectangular window comprises:
determining the confidence value of the central pixel point located at the central position of the rectangular window according to the following formula:

M
=

∑

i
,

j
=
1

N

⁢

⁢

P
ij

*

Q
ij

,

wherein
M represents the confidence value of the central pixel point at the central position of the rectangular window, P ij represents an element located in row i and column j in a first matrix, and Q ij represents an element located in row i and column j in a second matrix, wherein the first matrix is a preset N×N matrix, elements in each row in the first matrix are the same, and the element in row i and column j in the second matrix is a grayscale value of a pixel point in row i and column j in the rectangular window, wherein 1≤i≤N, 1≤j≤N, N is an odd number, and N is greater than or equal to 3.

4. The method of claim 3 , wherein an expression of the first matrix is:

P
=

[

n
1

n
1

n
1

n
0

n
0

n
0

n
1

n
1

n
1

]

,

wherein
P is the first matrix, and n o >n 1 ; or

P
=

[

n
1

n

Assignees

• Honor Device Co Ltd

Inventors

• Hu Hongwei
• Dong Chen
• Ding Xin
• Gao Wenmei

Classifications

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G06T3/40

  Scaling of whole images or parts thereof, e.g. expanding or contracting · CPC title

• G06T3/60

  Rotation of whole images or parts thereof · CPC title

• G06V40/161

  Detection; Localisation; Normalisation · CPC title

• G06V40/168

  Feature extraction; Face representation · CPC title