CN110599553B - A kind of skin color extraction and detection method based on YCbCr - Google Patents

A kind of skin color extraction and detection method based on YCbCr Download PDF

Info

Publication number
CN110599553B
CN110599553B CN201910851662.2A CN201910851662A CN110599553B CN 110599553 B CN110599553 B CN 110599553B CN 201910851662 A CN201910851662 A CN 201910851662A CN 110599553 B CN110599553 B CN 110599553B
Authority
CN
China
Prior art keywords
skin color
ycbcr
image
component
color space
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910851662.2A
Other languages
Chinese (zh)
Other versions
CN110599553A (en
Inventor
高美凤
付天豪
于力革
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangnan University
Original Assignee
Jiangnan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiangnan University filed Critical Jiangnan University
Priority to CN201910851662.2A priority Critical patent/CN110599553B/en
Publication of CN110599553A publication Critical patent/CN110599553A/en
Application granted granted Critical
Publication of CN110599553B publication Critical patent/CN110599553B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20028—Bilateral filtering

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Color Image Communication Systems (AREA)
  • Image Processing (AREA)
  • Processing Of Color Television Signals (AREA)

Abstract

本发明公开了一种基于YCbCr的肤色提取及检测方法，属于图像处理技术领域。所述方法根据亮度分量Y的范围采用不同公式提取Cr、Cb分量以实现对图像肤色区域的提取，对于Y≤20或者Y≥180，采用改进的由RGB颜色空间到YCbCr颜色空间转换公式获得Cr、Cb分量，使得强光和弱光环境下对于肤色的分割依然是清楚的，提高了肤色识别的鲁棒性，有效解决了强光和弱光环境下无法实现对图像肤色区域的提取的问题，使得肤色检测技术在图像处理技术领域的应用中能够更加准确的识别出人体肤色区域。

Figure 201910851662

The invention discloses a skin color extraction and detection method based on YCbCr, belonging to the technical field of image processing. The method uses different formulas to extract Cr and Cb components according to the range of the luminance component Y to realize the extraction of the skin color area of the image. For Y≤20 or Y≥180, an improved conversion formula from RGB color space to YCbCr color space is used to obtain Cr. , Cb component, so that the segmentation of skin color in strong light and weak light environment is still clear, which improves the robustness of skin color recognition, and effectively solves the problem that the extraction of image skin color area cannot be realized in strong light and weak light environment. , so that the skin color detection technology can more accurately identify the skin color area of the human body in the application of the image processing technology field.

Figure 201910851662

Description

Skin color extraction and detection method based on YCbCr
Technical Field
The invention relates to a skin color extraction and detection method based on YCbCr, belonging to the technical field of image processing.
Background
Skin color extraction is the basis of skin color detection, while skin color detection refers to the process of selecting a pixel region corresponding to human skin in an image, and with the rapid development of image processing technology, skin color detection technology is widely applied to face detection and recognition, face tracking, facial expression recognition, gesture recognition, man-machine interaction, image retrieval based on image content, video monitoring and the like, but image information acquired by a camera is constantly influenced by complex environments, in the case of a captured image with a large variation in illumination intensity, it is difficult to capture skin color information, therefore, the specific meaning expressed by the human body is difficult to identify, and in order to meet the requirement that people can still effectively identify the skin color of the human body under different illumination intensity environments, skin color information in the image needs to be effectively extracted to highlight the skin color area of the human body, so that the meaning to be expressed is better expressed.
In recent years, segmentation of skin color of images using the YCbCr color space, i.e., YUV, where "Y" represents brightness, i.e., a gray scale value, has been a research focus; and "U" and "V" denote chromaticity, which is used to specify the color of the pixel. The Y component is created through the RGB input signal by superimposing specific parts of the RGB signal together. "chroma" defines the hue and saturation of a color, represented by Cr and Cb, respectively. Because the skin color pixel points are distributed in an approximately elliptical area range on the plane of the YCbCr color space, skin color targets such as human faces, gestures and the like can be effectively segmented.
The color spaces adopted by the images based on skin color segmentation at present mainly comprise YCbCr color spaces and HSV color spaces, but the good effect is not achieved on the processing under the environments of strong light and weak light.
Disclosure of Invention
The invention provides a skin color extraction and detection method based on YCbCr, aiming at solving the problems that skin color targets such as human faces, gestures and the like cannot be effectively segmented under the strong light and weak light environment and the robustness of extracting a skin color area is not high at present.
The invention provides a YCbCr-based skin color extraction method, which adopts different formulas to extract Cr and Cb components according to the range of a brightness component Y after obtaining the brightness component Y of an image so as to realize the extraction of a skin color area of the image, and comprises the following steps:
when Y is less than or equal to 20 or Y is greater than or equal to 180, Cr and Cb components are obtained by adopting an improved conversion formula from RGB color space to YCbCr color space;
when 20< Y <180, Cr and Cb components are obtained by using YCbCr color space conversion formula recommended by international standard ITU-R BT.601.
Optionally, when Y is less than or equal to 20 or Y is greater than or equal to 180, the improved conversion formula from the RGB color space to the YCbCr color space is used to obtain the Cr and Cb components, and the improved conversion formula from the RGB color space to the YCbCr color space is:
Figure BDA0002197154330000021
Figure BDA0002197154330000022
wherein, theta0、θ1、λ1、λ0Is a parameter, θ, determined by the variation of different illumination0、θ1、λ1、λ0The solving formula of (2) is as follows:
Figure BDA0002197154330000023
Figure BDA0002197154330000024
Figure BDA0002197154330000025
Figure BDA0002197154330000026
wherein: y is(Cr)iIs to determine a certain brightness YiThe logarithm of the Cr component of the image under component, obtained by the YCbCr color space conversion formula used in International Standard ITU-R BT.601, is obtained
Figure BDA00021971543300000210
x(Cr)iIs the same brightness YiThe component being reciprocal, i.e.
Figure BDA0002197154330000027
y(Cb)iIs to determine the brightness YiYCbCr color space conversion in component down images by using international standard ITU-R BT.601The Cb component being a negative logarithm of the value, i.e. obtained by the formula
Figure BDA0002197154330000028
x(Cb)iIs the same brightness YiThe component being reciprocal, i.e.
Figure BDA0002197154330000029
m is obtaining different brightness YiNumber of sample skin tone pictures.
Optionally, when 20< Y <180, the YCbCr color space conversion formula used by the international standard ITU-R bt.601 is adopted to obtain the Cr and Cb components, and the YCbCr color space conversion formula used by the international standard ITU-R bt.601 is
Cb＝-0.148R-0.291G+0.439B+128
Cr＝0.439R-0.368G-0.071B+128
R, G, B are original pixel values in RGB space.
Optionally, the obtaining the luminance component Y of the image includes:
and acquiring an original skin color image, and filtering the original skin color image to ensure that the pixel weight with higher similarity degree in the image is higher, the edge is more obvious, the contrast is higher, and the smooth image is obtained.
Optionally, a bilateral filter is used for filtering the original skin color map, and a transfer function expression of the bilateral filter is as follows:
Figure BDA0002197154330000031
wherein the weighting factor w (i, j, k, l) depends on the product of the domain kernel d (i, j, k, l) and the value domain kernel r (i, j, k, l), i.e.:
Figure BDA0002197154330000032
Figure BDA0002197154330000033
Figure BDA0002197154330000034
wherein f represents an original skin color map, g (i, j) represents a processed image, (i, j) represents a position of a certain pixel point to be processed in g (i, j), f (k, l) represents a gray value of a neighborhood pixel of the pixel point at the position (i, j) in f, and σ represents the gray value of a neighborhood pixel of the pixel point at the position (i, j) in fdIs a smoothing coefficient, σ, defining a domain kernelrIs the smoothing parameter of the value domain kernel, and e is the euler constant.
The second purpose of the invention is to provide a skin color detection method based on YCbCr, which adopts the skin color extraction method based on YCbCr to extract skin color.
Optionally, the method further includes:
and performing Gaussian filtering on the extracted Cr and Cb components, reducing the spatial change of the Cr and Cb component pixels, weakening random noise, and performing image binarization by adopting a maximum inter-class variance method to obtain a skin-color binarization picture.
Optionally, the kernel function of the gaussian filtering is as follows:
Figure BDA0002197154330000035
wherein the vector
Figure BDA0002197154330000036
Representing the mask coordinate (x, y) point,
Figure BDA0002197154330000037
is the mask center point coordinate, i.e., the mean of the mask coordinates.
Optionally, a high-definition camera is used for acquiring the original skin color image.
The third purpose of the present invention is to provide an application of the YCbCr-based skin color extraction method and/or the YCbCr-based skin color detection method in the technical field of image processing.
The invention has the beneficial effects that:
after the brightness component Y of the image is obtained, Cr and Cb components are extracted by adopting different formulas according to the range of the brightness component Y so as to realize the extraction of the skin color area of the image, thereby effectively improving the robustness of extracting the skin color area under different illumination intensities; specifically, when Y is less than or equal to 20 or Y is greater than or equal to 180, Cr and Cb components are obtained by adopting an improved conversion formula from an RGB color space to a YCbCr color space; when 20< Y <180, obtaining Cr and Cb components by adopting a YCbCr color space conversion formula used in the international standard ITU-R BT.601; aiming at the characteristic that the image is fuzzy under the strong light and weak light environments, the invention realizes the segmentation of the skin color by adopting different modes according to the brightness separation range, improves the robustness of the skin color identification, effectively solves the problem that the extraction of the image skin color area can not be realized under the strong light and weak light environments, and ensures that the subsequent application in face detection and identification, face tracking, facial expression identification, gesture identification, man-machine interaction, image retrieval based on image content and video monitoring can more accurately identify the human body skin color area.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a flowchart of a skin color detection method based on YCbCr according to the present application.
FIG. 2 is a skin color original in the range of 150. ltoreq. Y.ltoreq.190.
Fig. 3 is a skin color map after bilateral filtering of fig. 2 using international standard ITU-R bt.601.
FIG. 4 is a gray scale and skin color combination diagram of Cb and Cr obtained by using skin color detection formula of Y in the range of 150 ≤ Y ≤ 190 of International Standard ITU-R BT.601.
Fig. 5 is a diagram after binarization in step (4).
FIG. 6 is a skin color original in the range of 0. ltoreq. Y.ltoreq.20.
Fig. 7 is a skin tone map for bilateral filtering of fig. 6 using international standard ITU-R bt.601.
Fig. 8 is a gray scale skin color combination plot of Cb, Cr obtained from fig. 7 using the improved skin color detection formula given in this application with Y in the range less than 20.
Fig. 9 is a grey scale skin color combination map of Cb, Cr obtained from fig. 7 using the YCbCr color space conversion formula used by international standard ITU-R bt.601.
Fig. 10 is a diagram of fig. 8 after the binarization at step (4).
Fig. 11 is a diagram of fig. 9 after the binarization at step (4).
FIG. 12 is a skin color original in the range of 180 ≦ Y ≦ 200.
Fig. 13 is a skin tone map for bilateral filtering of fig. 12 using international standard ITU-R bt.601.
FIG. 14 is a combined graph of the gray level skin color of Cb and Cr obtained by using the skin color detection formula given in the international standard ITU-R BT.601 in the range of Y being more than or equal to 180 and less than or equal to 200.
Fig. 15 is a diagram of fig. 14 after the binarization at step (4).
FIG. 16 is a gray scale skin color combination graph of Cb and Cr obtained by using skin color detection formula with the range of Y being more than or equal to 180 and less than or equal to 200 provided by the application
Fig. 17 is a diagram of fig. 16 after the binarization at step (4).
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
The first embodiment is as follows:
the embodiment provides a skin color detection method based on YCbCr, which comprises the steps of converting an RGB color space into an YCbCr color space, obtaining a brightness component of an image, and extracting Cr and Cb components by adopting different formulas according to the range of the brightness component Y so as to extract a skin color area of the image;
the following example of the extraction of skin color blocks under different illumination intensities illustrates the specific implementation of the method of the present invention, and the experimental environment includes: the CPU adopts i5-7200U, adopts Python language to write programs, processes images in a Spyder program debugging environment, and adopts a Window10 operating system in a system running environment. The whole improved skin color segmentation method hardware part experimental environment comprises the following steps: the camera adopts Roots (Logitech) C270iIPTV high-definition network camera, adopts an LED annular light source of DHO-RI12030 of great Heng group Limited company in China to adjust the brightness, and the invention is further described with reference to the attached drawings, and please refer to FIG. 1:
step 1: acquiring an original skin color image, and filtering the original skin color image;
in this embodiment, the skin color original image (i.e., fig. 2) with Y being greater than or equal to 150 and less than or equal to 190, the skin color original image (i.e., fig. 6) with Y being greater than or equal to 0 and less than or equal to 20, and the skin color original image (i.e., fig. 12) with Y being greater than or equal to 180 and less than or equal to 200 are selected for extracting skin color.
Filtering the original skin color image to ensure that the pixel weight with higher similarity degree in the image is higher, the edge is more obvious and the contrast is higher:
specifically, a bilateral filter is used for image filtering, a software program calls a bilatelfilter () function of a cv2 module of the OpenCV, and the bilateral filter is used for image filtering, wherein the diameter of a pixel to be filtered is 5 pixels, and the result graphs after filtering in fig. 2, 6, and 12 are respectively shown in fig. 3, 7, and 13, and the transfer function expression of the bilateral filter is as follows:
Figure BDA0002197154330000051
wherein the weighting factor w (i, j, k, l) depends on the product of the domain kernel d (i, j, k, l) and the value domain kernel r (i, j, k, l), i.e.:
Figure BDA0002197154330000052
Figure BDA0002197154330000053
Figure BDA0002197154330000054
wherein f represents an original skin color map, g (i, j) represents a processed image, (i, j) represents a position of a certain pixel point to be processed in g (i, j), f (k, l) represents a gray value of a neighborhood pixel of the pixel point at the position (i, j) in f, and σ represents the gray value of a neighborhood pixel of the pixel point at the position (i, j) in fdIs a smoothing coefficient, σ, defining the kernel of the domainrIs the smoothing parameter of the value domain kernel, and e is the euler constant.
Step 2: aiming at the images 3, 7 and 13 obtained by the bilateral filtering in the step 1, the conversion of the RGB color space and the YCbCr color space is carried out to obtain the converted brightness component, and the conversion formula is as follows:
Y＝0.257R+0.504G+0.098B+16
r, G, B are original pixel values in RGB space, and Y is pixel value of YCbCr space brightness component.
And step 3: judging the approximate range of the luminance Y component of the skin color picture after conversion in step 2, wherein the approximate range of the Y component in fig. 3 is as follows: y is more than or equal to 150 and less than or equal to 190, and for the skin color picture in the brightness component range, a YCbCr color space conversion formula recommended by the international standard ITU-R BT.601 in the prior art is adopted to obtain Cr and Cb components, wherein the conversion formula is as follows:
Cb＝-0.148R-0.291G+0.439B+128
Cr＝0.439R-0.368G-0.071B+128
obtaining a gray scale image of the combination of the Cr and Cb components after conversion, as shown in FIG. 4;
and the approximate range of the Y component of fig. 7 is: y is more than or equal to 0 and less than or equal to 20, Cr and Cb components are obtained by adopting an improved conversion formula from RGB color space to YCbCr color space, and a gray scale image combining the Cr and Cb components is obtained after conversion, as shown in FIG. 8; wherein the conversion formula is as follows:
Figure BDA0002197154330000061
Figure BDA0002197154330000062
wherein, theta0、θ1、λ1、λ0Is a parameter, θ, determined by the variation of different illumination0、θ1、λ1、λ0The solving formula of (2) is as follows:
Figure BDA0002197154330000063
Figure BDA0002197154330000064
Figure BDA0002197154330000065
Figure BDA0002197154330000066
wherein: y is(Cr)iIs to determine a certain brightness YiThe logarithm of the Cr component in the component-down image obtained by the YCbCr color space conversion formula used in International Standard ITU-R BT.601 in step (3), i.e. the value obtained by taking the logarithm of the Cr component
Figure BDA0002197154330000067
Figure BDA0002197154330000068
x(Cr)iIs the same brightness YiThe component being reciprocal, i.e.
Figure BDA0002197154330000069
y(Cb)iIs to determine the brightness YiThe color space conversion formula of YCbCr used in international standard ITU-R BT.601 in step (3) in the image under componentThe Cb component taking a logarithmic value, i.e.
Figure BDA00021971543300000610
x(Cb)iIs the same brightness YiThe component being reciprocal, i.e.
Figure BDA0002197154330000071
m is obtaining different brightness YiNumber of sample skin tone pictures.
The approximate range of the Y component of fig. 12 is: y is more than or equal to 180 and less than or equal to 200, Cr and Cb component combined graphs are obtained by respectively adopting a conversion formula from an RGB color space to a YCbCr color space of a YCbCr color space conversion formula used in the international standard ITU-R BT.601, as shown in figure 14, and a binary graph of Cr and Cb component combined is obtained after conversion, as shown in figure 15; and the improved conversion formula from RGB color space to YCbCr color space herein is to obtain the Cr, Cb component combination map as shown in fig. 16, and obtain the binarized map after conversion of Cr and Cb components as shown in fig. 17, while if the conversion formula from YCbCr color space, which is not optimized in international standard ITU-R bt.601, is used in fig. 7 to obtain the gray scale map obtained by combining Cr, Cb components, and the obtained gray scale map is shown in fig. 9, it is known that the improved conversion formula from RGB color space to YCbCr color space provided by the present application has a good effect of extracting the skin color picture with Y components in the range of 0 ≦ Y ≦ 20 compared to the gray scale map obtained by combining Cr, Cb components using the improved conversion formula from RGB color space to YCbCr color space provided by the present application, i.e. fig. 8. As can be seen from comparison of FIG. 15 and FIG. 17, it is also effective in extracting skin color under strong light 180. ltoreq. Y.ltoreq.200.
In this embodiment, m is 600, and one group of parameters obtained and adopted from the skin color sample picture under the condition of 600 different brightnesses is: theta0＝2.92；θ1＝-52.07，λ0＝2.42，λ1＝-28.27。
And (4) performing Gaussian filtering on the gray level image 4, the gray level image 8 and the gray level image 9 combined by the Cr and Cb components in the step (3), and performing binarization on the image to obtain a picture after skin color binarization, wherein the obtained result is shown as a binarization image 5, a binarization image 10 and a binarization image 11. Wherein the gaussian filter kernel function is as follows:
Figure BDA0002197154330000072
wherein the vector
Figure BDA0002197154330000073
Representing the mask coordinate (x, y) point,
Figure BDA0002197154330000074
is the mask center point coordinate, i.e., the mean of the mask coordinates.
Fig. 10 is a result diagram after binarization in fig. 8, fig. 11 is a result diagram after binarization in fig. 9, and comparing fig. 10 and fig. 11 with comparing fig. 15 and fig. 17, it can be known that in the skin color detection method based on YCbCr provided in the present application, the Cr and Cb components are obtained by an improved conversion formula from an RGB color space to an YCbCr color space, and then a gray scale image in which the Cr and Cb components are combined is obtained, so that robustness of extracting skin color regions under different illumination intensities is effectively improved, and a segmentation effect is significantly improved.
Some steps in the embodiments of the present invention may be implemented by software, and the corresponding software program may be stored in a readable storage medium, such as an optical disc or a hard disk.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (10)

1. A skin color extraction method based on YCbCr is characterized in that after a brightness component Y of an image is obtained, Cr and Cb components are extracted by adopting different formulas according to the range of the brightness component Y so as to realize the extraction of a skin color area of the image, and the method comprises the following steps:
when Y is less than or equal to 20 or Y is greater than or equal to 180, Cr and Cb components are obtained by adopting an improved conversion formula from RGB color space to YCbCr color space;
when Y is more than 20 and less than 180, obtaining Cr and Cb components by adopting a YCbCr color space conversion formula used by the international standard ITU-R BT.601;
the improved conversion formula from the RGB color space to the YCbCr color space is as follows:
Figure FDA0003149238060000011
Figure FDA0003149238060000012
wherein, theta0、θ1、λ1、λ0Is a parameter, θ, determined by the variation of different illumination0、θ1、λ1、λ0The solving formula of (2) is as follows:
Figure FDA0003149238060000013
Figure FDA0003149238060000014
Figure FDA0003149238060000015
Figure FDA0003149238060000016
wherein: y is(Cr)iIs to determine a certain brightness YiThe logarithm of the Cr component of the image under component, obtained by the YCbCr color space conversion formula used in International Standard ITU-R BT.601, is obtained
Figure FDA0003149238060000017
x(Cr)iIs the same brightness YiThe component being reciprocal, i.e.
Figure FDA0003149238060000018
y(Cb)iIs to determine the brightness YiThe Cb component obtained by YCbCr color space conversion formula used by International Standard ITU-R BT.601 in the image under the component is a value obtained by taking the negative logarithm, namely
Figure FDA0003149238060000019
x(Cb)iIs the same brightness YiThe component being reciprocal, i.e.
Figure FDA00031492380600000110
Figure FDA00031492380600000111
m is obtaining different brightness YiNumber of sample skin tone pictures.
2. The method of claim 1, wherein when 20< Y <180, the Cr and Cb components are obtained by using YCbCr color space conversion formula used in international standard ITU-R bt.601, and the YCbCr color space conversion formula used in international standard ITU-R bt.601 is
Cb＝-0.148R-0.291G+0.439B+128
Cr＝0.439R-0.368G-0.071B+128
R, G, B are original pixel values in RGB space.
3. The method of claim 2, wherein obtaining the luminance component Y of the image comprises:
and acquiring an original skin color image, and filtering the original skin color image to ensure that the pixel weight with higher similarity degree in the image is higher, the edge is more obvious, the contrast is higher, and the smooth image is obtained.
4. The method of claim 3, wherein the filtering of the raw skin map employs a bilateral filter whose transfer function is expressed as follows:
Figure FDA0003149238060000021
wherein the weighting factor w (i, j, k, l) depends on the product of the domain kernel d (i, j, k, l) and the value domain kernel r (i, j, k, l), i.e.:
Figure FDA0003149238060000022
Figure FDA0003149238060000023
Figure FDA0003149238060000024
wherein f represents an original skin color map, g (i, j) represents a processed image, (i, j) represents a position of a certain pixel point to be processed in g (i, j), f (k, l) represents a gray value of a neighborhood pixel of the pixel point at the position (i, j) in f, and σ represents the gray value of a neighborhood pixel of the pixel point at the position (i, j) in fdIs a smoothing coefficient, σ, defining a domain kernelrIs the smoothing parameter of the value domain kernel, and e is the euler constant.
5. A YCbCr-based skin color detection method, characterized in that the YCbCr-based skin color detection method adopts the method of any one of claims 1-4 to extract skin color.
6. The method of claim 5, further comprising:
and performing Gaussian filtering on the extracted Cr and Cb components, reducing the spatial change of the Cr and Cb component pixels, weakening random noise, and performing image binarization by adopting a maximum inter-class variance method to obtain a skin-color binarization picture.
7. The method of claim 6, wherein the kernel function of the Gaussian filter is as follows:
Figure FDA0003149238060000025
wherein the vector
Figure FDA0003149238060000031
Representing the mask coordinate (x, y) point,
Figure FDA0003149238060000032
is the mask center point coordinate, i.e., the mean of the mask coordinates.
8. The YCbCr-based skin color extraction method according to any of the claims 1-4, wherein the original skin color image is obtained by using a high-definition camera.
9. The application of a skin color extraction method in the technical field of image processing is characterized in that the skin color extraction method is the YCbCr-based skin color extraction method according to any one of claims 1 to 4.
10. Use of a skin color detection method in the field of image processing technology, wherein the skin color detection method is the YCbCr-based skin color detection method according to any one of claims 6 to 9.
CN201910851662.2A 2019-09-10 2019-09-10 A kind of skin color extraction and detection method based on YCbCr Active CN110599553B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910851662.2A CN110599553B (en) 2019-09-10 2019-09-10 A kind of skin color extraction and detection method based on YCbCr

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910851662.2A CN110599553B (en) 2019-09-10 2019-09-10 A kind of skin color extraction and detection method based on YCbCr

Publications (2)

Publication Number Publication Date
CN110599553A CN110599553A (en) 2019-12-20
CN110599553B true CN110599553B (en) 2021-11-02

Family

ID=68858558

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910851662.2A Active CN110599553B (en) 2019-09-10 2019-09-10 A kind of skin color extraction and detection method based on YCbCr

Country Status (1)

Country Link
CN (1) CN110599553B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113642358B (en) * 2020-04-27 2023-10-10 华为技术有限公司 Skin color detection method, device, terminal and storage medium
CN112991366B (en) * 2021-02-18 2024-05-03 广州光锥元信息科技有限公司 Method, device and mobile terminal for carrying out real-time chromaticity matting on image
CN115331286B (en) * 2022-07-29 2023-03-17 中国兵器工业信息中心 Content safety detection system based on deep learning

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050050061A (en) * 2005-04-30 2005-05-27 조동욱 Facial features extraction for ocular inspection
CN103268472A (en) * 2013-04-17 2013-08-28 哈尔滨工业大学深圳研究生院 Lip detection method based on dual color space
CN104050488A (en) * 2014-06-16 2014-09-17 西安工业大学 Hand gesture recognition method based on switching Kalman filtering model
CN104679242A (en) * 2015-02-05 2015-06-03 吉林纪元时空动漫游戏科技股份有限公司 Hand gesture segmentation method based on monocular vision complicated background
CN104715244A (en) * 2015-04-01 2015-06-17 华中科技大学 Multi-viewing-angle face detection method based on skin color segmentation and machine learning
CN108805097A (en) * 2018-06-21 2018-11-13 冷霜 A kind of recognition of face and method for real time tracking based on color space conversion
KR20190056750A (en) * 2017-11-17 2019-05-27 김영수 Apparatus and method for detecting by using image process
CN109934152A (en) * 2019-03-08 2019-06-25 浙江理工大学 An improved small curved arm image segmentation method for sign language images

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5032911B2 (en) * 2007-07-31 2012-09-26 キヤノン株式会社 Image processing apparatus and image processing method

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050050061A (en) * 2005-04-30 2005-05-27 조동욱 Facial features extraction for ocular inspection
CN103268472A (en) * 2013-04-17 2013-08-28 哈尔滨工业大学深圳研究生院 Lip detection method based on dual color space
CN104050488A (en) * 2014-06-16 2014-09-17 西安工业大学 Hand gesture recognition method based on switching Kalman filtering model
CN104679242A (en) * 2015-02-05 2015-06-03 吉林纪元时空动漫游戏科技股份有限公司 Hand gesture segmentation method based on monocular vision complicated background
CN104715244A (en) * 2015-04-01 2015-06-17 华中科技大学 Multi-viewing-angle face detection method based on skin color segmentation and machine learning
KR20190056750A (en) * 2017-11-17 2019-05-27 김영수 Apparatus and method for detecting by using image process
CN108805097A (en) * 2018-06-21 2018-11-13 冷霜 A kind of recognition of face and method for real time tracking based on color space conversion
CN109934152A (en) * 2019-03-08 2019-06-25 浙江理工大学 An improved small curved arm image segmentation method for sign language images

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Face Detection Method Based on Multi-feature Fusion in YCbCr Color Space;Youlian Zhu 等;《2012 5th International Congress on Image and Signal Processing》;20121231;第1249-1252页 *
基于YCbCr高斯肤色模型的人脸检测技术研究;王航宇;《现代电子技术》;20081231;第102-105页 *

Also Published As

Publication number Publication date
CN110599553A (en) 2019-12-20

Similar Documents

Publication Publication Date Title
Silva et al. Near real-time shadow detection and removal in aerial motion imagery application
US8358813B2 (en) Image preprocessing
US8559708B2 (en) Image preprocessing
US8660323B2 (en) Image Preprocessing
Ghazali et al. An innovative face detection based on skin color segmentation
CN110599553B (en) A kind of skin color extraction and detection method based on YCbCr
CN107220624A (en) A kind of method for detecting human face based on Adaboost algorithm
CN110428439A (en) A kind of shadow detection method based on shadow region color saturation property
Lee et al. Color image enhancement using histogram equalization method without changing hue and saturation
US8244004B2 (en) Image preprocessing
CN110458012B (en) Multi-angle face recognition method and device, storage medium and terminal
CN101364304A (en) Shadow detection method based on color invariance and Gaussian model
Meng et al. Dominant points based hand finger counting for recognition under skin color extraction in hand gesture control system
Yusuf et al. Human face detection using skin color segmentation and watershed algorithm
Kumar et al. Real-time human skin color detection algorithm using skin color map
CN108133204B (en) Hand body identification method, device, equipment and computer readable storage medium
Manjare et al. Skin detection for face recognition based on HSV color space
CN105719304B (en) A kind of flower image dividing method based on Otsu
Liang et al. Applying Image Processing Technology to Face Recognition.
CN108549855A (en) Real-time humanoid detection method towards Intelligent household scene
Abdallah et al. Different techniques of hand segmentation in the real time
Ojo et al. Illumination invariant face detection using hybrid skin segmentation method
Vincent et al. Shadow detection: A review of various approaches to enhance image quality
Higashi et al. New feature for shadow detection by combination of two features robust to illumination changes
Ma'asum et al. Comparative study of segmentation technique for hand gesture recognition system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant