Speckle detection method and electronic device

Patent metadata

Field	Value
Publication number	US-11989885-B2
Application number	US-201817260855-A
Country	US
Kind code	B2
Filing date	Sep 18, 2018
Priority date	Jul 16, 2018
Publication date	May 21, 2024
Grant date	May 21, 2024

Abstract

A speckle detection method includes: obtaining a to-be-detected image; converting the to-be-detected image into Lab color space to obtain a Lab image; extracting a speckle feature from the Lab image to obtain a speckle feature image, where the speckle feature image includes a skin speckle feature and a subcutaneous speckle feature; and determining a skin speckle and a subcutaneous speckle in the speckle feature image.

First claim

What is claimed is:

1. A speckle detection method, comprises:
obtaining a to-be-detected image;
converting the to-be-detected image into Lab color space to obtain a Lab image;
extracting a speckle feature from the Lab image to obtain a speckle feature image, comprising separately extracting detail feature components of a channel L, a channel a, and a channel b in the Lab image to obtain the speckle feature image, wherein the speckle feature image comprises a skin speckle feature and a subcutaneous speckle feature; and
determining a skin speckle and a subcutaneous speckle in the speckle feature image, comprising
determining one or more speckle regions in the speckle feature image,
for each of the one or more speckle regions, determining a channel b average value of pixel points of the respective speckle region, determining the respective speckle region whose channel b average value is greater than a first threshold as the skin speckle, and determining the respective speckle region whose channel b average value is less than or equal to the first threshold as the subcutaneous speckle.

2. The method of claim 1 , extracting the speckle feature to obtain the speckle feature image further comprises:
determining a channel L difference between the channel L and the extracted detail feature component of the channel L, a channel a difference between the channel a and the extracted detail feature component of the channel a, and a channel b difference between the channel b and the extracted detail feature component of the channel b; and
obtaining the speckle feature image based on the channel L difference, the channel a difference, and the channel b difference, wherein the channel L in the speckle feature image is the channel L difference, the channel a in the speckle feature image is the channel a difference, and the channel b in the speckle feature image is the channel b difference.

3. The method of claim 2 , wherein the method further comprises:
separately performing bilateral filtering processing on the channel L, the channel a, and the channel b in the Lab image, to obtain the detail feature component of the channel L, the detail feature component of the channel a, and the detail feature component of the channel b.

4. The method of claim 2 , wherein determining the one or more speckle regions further comprises:
determining a first pixel point in a detection frame, wherein a length of the detection frame is less than a length of the speckle feature image, a width of the detection frame is less than a width of the speckle feature image, the detection frame moves in the speckle feature image at a preset step, and a pixel value of the first pixel point satisfies the following formula:
r 1<( a 1− T 1× b 1), wherein

r1 is the pixel value of the first pixel point, a1 is an average value of pixel values of pixel points in the detection frame, T1 is a preset value, and b1 is a variance of the pixel values of the pixel points in the detection frame;
determining a second pixel point in the speckle feature image, wherein a pixel value of the second pixel point satisfies the following formula:
r 2<( a 2− T 2× b 2), wherein

r2 is the pixel value of the second pixel point, a2 is an average value of pixel values of pixel points in the speckle feature image, T2 is a preset value, and b2 is a variance of the pixel values of the pixel points in the speckle feature image;
determining the first pixel point and the second pixel point as speckles; and
performing an expansion operation on one of the speckles, and performing a corrosion operation on that speckle on which the expansion operation is performed, to obtain the speckle region.

5. The method of claim 2 , wherein the method further comprises:
removing the respective speckle region whose area is less than a second threshold and/or whose area is greater than a third threshold, wherein the second threshold is less than the third threshold; and/or
removing the respective speckle region whose ratio of an area to a circumference is less than a fourth threshold.

6. The method of claim 1 , wherein the method further comprises:
determining a first feature set, and quantizing a score of the skin speckle based on the first feature set, wherein the first feature set comprises at least one of the following features: a uniformity value, a quantity of skin speckles, a speckle area of the skin speckle, or a contrast value of the skin speckle, the uniformity value is used to represent pigment uniformity of the speckle feature image, and the contrast value of the skin speckle is used to represent color contrast of the skin speckle;
determining a second feature set, and quantizing a score of the subcutaneous speckle based on the second feature set, wherein the second feature set comprises at least one of the following features: the uniformity value, a quantity of subcutaneous speckles, a speckle area of the subcutaneous speckle, or a contrast value of the subcutaneous speckle, and the contrast value of the subcutaneous speckle is used to represent color contrast of the subcutaneous speckle;
determining a comprehensive score of speckle detection based on the score of the skin speckle and the score of the subcutaneous speckle; and
displaying the comprehensive score, or displaying the score of the skin speckle, the score of the subcutaneous speckle, and the comprehensive score.

7. The method of claim 6 , wherein the score of the skin speckle is determined according to the following formula:

H
1

=

w
1

×
A

+

w
2

×

B
1

+

w
3

×

C
1

×

D
1

E

,

 wherein
H 1 is the score of the skin speckle, A is the uniformity value, B 1 is the quantity of skin speckles, C 1 is a sum of contrast values of all the skin speckles, D 1 is a sum of areas of all the skin speckles, E is an area of the speckle feature image, and w 1 , w 2 , and w 3 are all preset parameters;
the score of the subcutaneous speckle is determined according to the following formula:

H
2

=

w
1

×
A

+

w
4

×

B
2

Assignees

• Honor Device Co Ltd

Inventors

• Guo Zhizhi
• Lu Henghui
• Gao Wenmei

Classifications

• G06T7/11 Primary

  Region-based segmentation · CPC title

• G06F18/241

  relating to the classification model, e.g. parametric or non-parametric approaches · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T7/514

  from specularities · CPC title

• G06V10/56

  relating to colour · CPC title