CN112767257A - Method for improving accuracy of reduction of different skin colors - Google Patents
Method for improving accuracy of reduction of different skin colors Download PDFInfo
- Publication number
- CN112767257A CN112767257A CN202011427655.9A CN202011427655A CN112767257A CN 112767257 A CN112767257 A CN 112767257A CN 202011427655 A CN202011427655 A CN 202011427655A CN 112767257 A CN112767257 A CN 112767257A
- Authority
- CN
- China
- Prior art keywords
- color
- skin
- skin color
- reduction
- accuracy
- 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.)
- Pending
Links
- 239000003086 colorant Substances 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 17
- 241000282414 Homo sapiens Species 0.000 claims abstract description 32
- 238000005286 illumination Methods 0.000 claims abstract description 28
- 239000011159 matrix material Substances 0.000 claims abstract description 23
- 230000000694 effects Effects 0.000 claims abstract description 8
- 238000007405 data analysis Methods 0.000 claims abstract description 6
- 206010048245 Yellow skin Diseases 0.000 description 8
- 201000005569 Gout Diseases 0.000 description 2
- 241001125929 Trisopterus luscus Species 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001085205 Prenanthella exigua Species 0.000 description 1
- 206010040829 Skin discolouration Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
According to the method for improving the accuracy of the reduction of different skin colors, provided by the invention, through big data analysis, 4 skin colors of human beings all over the world are deeply divided into yellow, white, black and brown to prepare a skin color 24 color card representing the skin colors of the human beings all over the world; shooting raw data original images of the 24-color card with skin color under different standard light sources and different illumination intensities to obtain calibration data; constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities; determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight; and obtaining the color temperature and the illumination of the current scene according to the sensor by adopting bilinear interpolation, then obtaining a corresponding color reduction matrix CCM by applying the bilinear interpolation, and acting on the whole image to obtain a live effect image which is closer to the skin color of the human face. The accuracy of restoring the skin color is better, and the accuracy of restoring the skin color in a complex environment is improved.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a method for improving the accuracy of different types of skin color reduction.
Background
With the development of the camera photographing technology in the society at present, people have higher and higher requirements on the accuracy of photographing color reduction, and people with 4 skin colors, yellow, white, black and brown, are on the earth. Actually, human skin colors are countless because human skin colors are ever-changing, and with the addition of different regional differences, different racial marriage, different working environments and the like, multicolor skins are formed, and especially the accuracy requirement on human face skin color reduction is higher and higher, but the current algorithm for human face skin color reduction is relatively single, the human face skin color reduction is enhanced, most of the human face skin color reduction is achieved by changing Rgain and Bgain global compensation gain, the influence on other colors is larger, and the accuracy of color reduction of other skin colors cannot be solved.
In summary, there is a need for a method capable of solving the above technical problems, in which a skin color 24 card representing the skin colors of the human beings all over the world as much as possible is further divided into 4 skin colors of the human beings all over the world, a face recognition algorithm is used to determine the approximate skin color of a photographed target, and then a CCM matrix is solved by giving higher weight to the corresponding skin color, so that the accuracy of target skin color reduction is further improved, and bilinear interpolation is adopted to achieve better skin color reduction accuracy, thereby improving the accuracy of complex environment skin color reduction and overcoming the above problems.
Disclosure of Invention
In view of the defects of the prior art, the invention aims to provide a method for improving the accuracy of restoring different types of skin colors, and aims to solve the problem that the accuracy of restoring the skin colors of human faces in the prior art is low.
In order to achieve the purpose, the invention adopts the following technical scheme:
a method for improving accuracy of reduction of different skin colors is characterized by comprising the following steps:
s10, carrying out further division on four current human skin colors (yellow, white, black and brown) all over the world through big data analysis to prepare a 24-color card of the skin color representing the human skin color all over the world;
s20, shooting raw data original images of the 24-color card with skin color under different illumination of different standard light sources, and acquiring calibration data, wherein the different standard light sources are as follows: d75, D65, D50, Tl84, CWF, U30, F, A, the different illumination intensities being: 1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux, 10 lux;
s30, constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities;
s40, determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight to the corresponding skin color when solving a color reduction matrix;
s50, obtaining the color temperature and the illumination of a scene according to a sensor, and then applying bilinear interpolation to obtain a corresponding color reduction matrix to act on the whole image to obtain a live effect image which is closer to the skin color of the human face.
Compared with the prior art, the invention has the beneficial effects that:
according to the method for improving the accuracy of the reduction of different skin colors, provided by the invention, through big data analysis, 4 skin colors of human beings all over the world at present are deeply divided into yellow, white, black and brown to prepare a 24-color skin color card which represents the skin colors of human beings all over the world as far as possible; under different standard light sources (D75, D65, D50, Tl84, CWF, U30 and F, A) and under different illumination intensities (1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux and 10lux), raw data original images of the skin color 24 color cards obtained in the steps are shot, and calibration data are obtained; constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities; determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight; the method comprises the steps of considering the influence of color temperature on color restoration and also considering the restoration of illumination on color, adopting bilinear interpolation, obtaining the color temperature and the illumination of the current scene according to a sensor, obtaining a corresponding color restoration matrix CCM by applying the bilinear interpolation, and acting on the whole image to obtain a live effect image which is closer to the complexion of the human face. The accuracy of restoring the skin color is better, and the accuracy of restoring the skin color in a complex environment is improved.
Drawings
FIG. 1 is a block flow diagram of the present invention.
Detailed Description
In order to make the objects, technical solutions and effects of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in the attached figure 1, the method for improving the accuracy of the reduction of different skin colors provided by the invention is characterized by comprising the following steps:
s10, carrying out further division on four current human skin colors (yellow, white, black and brown) all over the world through big data analysis to prepare a 24-color card of the skin color representing the human skin color all over the world;
s20, shooting raw data original images of the 24-color card with skin color under different illumination of different standard light sources, and acquiring calibration data, wherein the different standard light sources are as follows: d75, D65, D50, Tl84, CWF, U30, F, A, the different illumination intensities being: 1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux, 10 lux;
s30, constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities;
s40, determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight to the corresponding skin color when solving a color reduction matrix;
s50, obtaining the color temperature and the illumination of a scene according to a sensor, and then applying bilinear interpolation to obtain a corresponding color reduction matrix to act on the whole image to obtain a live effect image which is closer to the skin color of the human face.
Generally, there are roughly 4 people with skin color on earth, yellow, white, black, brown. In fact, human skin color is countless because human skin color is varied, and forms multicolor skin in addition to different regional differences, different racial marriage, different work environments, and the like. The invention is to cover common skin color as far as possible, and not break the integral structure of the original 24-color card, and the yellow skin color is divided into three types: normal yellow, dark yellow, reddish yellow; skin lightening is divided into three categories: normal white skin color, dark white skin color, bright white skin color; black skin colors are classified into three categories: normal black skin color, dark skin color, bright black skin color; brown skin colors are divided into three categories: normal brown skin, pale brown skin, dark brown skin.
Taking yellow skin as an example, a normal yellow skin color, a dark yellow skin color and a red yellow skin color (R, G, B) acquisition method is described below, 10000 yellow skin high and middle school face skin colors are randomly collected, and pixel values Pi (Ri, Gi, Bi) of the 10000 face skin colors are counted, wherein 1< ═ i < ═ 1000, and Pi (Ri, Gi, Bi) represents the ith high and middle school face skin color (R, G, B).
The pixel point values in the SRGB space of 12 skin colors under four skin colors are obtained by the method, and then the first 12 color blocks of the Alice 24 color card are sequentially replaced by the following 12 colors, so that a new skin color 24 color card template is obtained. The 12 colors can only replace the first 12 color blocks of the Alice standard 24 color card, because the first 12 colors of the standard 24 color card are colors in nature, and the 13-18 color blocks represent RGB three main colors and three auxiliary colors, are basic colors for color restoration and cannot be modified. However, the specific positions of the first 12 color blocks of the standard 24 color card where the 12 skin colors are located can be randomly ordered.
The table of RGB values for 12 skin tones is as follows:
and when one color temperature is calculated according to the AWB compensation gain, the CCM matrix of the current color temperature is obtained by performing linear interpolation on the CCM matrixes under the two color temperatures according to the color temperature section where the color temperature is located.
The method comprises the steps of shooting RAW data original images of 24-color cards of skin colors obtained in the steps under different illumination intensities (1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux and 10lux) under different standard light sources (D75, D65, D50, Tl84, CWF, U30 and F, A), obtaining 64 RAW data original images under 8 color temperatures and 8 illumination intensities, counting (R, G and B) of each color block, establishing a relation according to original (R, G and B) of the 24-color card color block of the skin colors, and obtaining a 3 x 3 conversion relation matrix of RGB three channels, namely a CCM matrix, wherein the CCM matrix enables the value of each block of the shot 24-color cards of the skin colors to be closer to the value of each color block of the 24-color cards of the original color blocks, and the 3 x 3 conversion relation matrix can be obtained through the following relation:
pin _ k (Rin _ k, Gin _ k, Bin _ k), Pout _ k (Rout _ k, Gout _ k, Bout _ k) where 1< ═ k < ═ 24, Pin _ k (Rin _ k, Gin _ k, Bin _ k) represents RGB values of the k-th patch of a raw image of a skin color 24 color card photographed by camera; pout _ k (Rout _ k, Gout _ k, Bout _ k) represents the RGB values of the k-th patch of the formulated skin color 24-patch image;
solving the CCM matrix becomes solving the following over-determined equation (the number of equations is greater than the number of unknowns):
solving the CCM matrix C of 3 × 3 minimizes the function value where Pk represents the weight value of the k-th patch, and if we want to make the yellow skin tone restoration more accurate, we can give a higher weight value to the yellow skin tone patch.
Thus, 64 groups of CCM matrixes under different illumination intensities (1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux and 10lux) under different standard light sources (D75, D65, D50, Tl84, CWF, U30 and F, A) are obtained.
According to a face recognition algorithm, determining that a shooting target is a face, then picking out a face region, and counting RGB values of the face region: (Rface, Gface, Bface), determining the skin color (yellow, white, black, brown) of the shot human face of the subject by the following method:
wherein, as "table: the RGB value of 12 skin colors (Rij, Gij, Bij) represents one of 12 skin colors, when a shot picture is identified to contain one skin color, a CCM color restoration matrix is calculated, a higher weight value is given to the skin color, when one skin color is identified, the CCM color restoration matrix is calculated, a higher weight value is given to the skin color, when multiple skin colors are identified, the CCM color restoration matrix is calculated, higher weight values are given to the skin colors, and if a human face is not identified, the CCM matrix calculated by using an original Alice standard color card is used. Implementation of this step for a camera, we can acquire the identified face skin color in a preview mode and then use the updated CCM matrix parameters in a photo mode.
Through the steps, 64 groups of C under different illumination intensities (1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux and 10lux) under different standard light sources (D75, D65, D50, Tl84, CWF, U30 and F, A) are obtained as CCM matrixes obtained according to different skin color weights; and then, obtaining the color temperature and the illumination of the current scene according to the sensor, then applying bilinear interpolation to obtain a corresponding color reduction matrix CCM, and acting on the whole image to obtain a live effect image which is closer to the skin color of the human face.
The technology can be applied to other electronic equipment with camera functions besides camera mobile phones for taking pictures.
In summary, the working principle of the invention is as follows:
according to the method for improving the accuracy of the reduction of different skin colors, provided by the invention, through big data analysis, 4 skin colors of human beings all over the world at present are deeply divided into yellow, white, black and brown to prepare a 24-color skin color card which represents the skin colors of human beings all over the world as far as possible; under different standard light sources (D75, D65, D50, Tl84, CWF, U30 and F, A) and under different illumination intensities (1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux and 10lux), raw data original images of the skin color 24 color cards obtained in the steps are shot, and calibration data are obtained; constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities; determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight; the method comprises the steps of considering the influence of color temperature on color restoration and also considering the restoration of illumination on color, adopting bilinear interpolation, obtaining the color temperature and the illumination of the current scene according to a sensor, obtaining a corresponding color restoration matrix CCM by applying the bilinear interpolation, and acting on the whole image to obtain a live effect image which is closer to the complexion of the human face. The accuracy of restoring the skin color is better, and the accuracy of restoring the skin color in a complex environment is improved.
It should be understood that equivalents and modifications of the technical solution and inventive concept thereof may occur to those skilled in the art, and all such modifications and alterations should fall within the scope of the appended claims.
Claims (1)
1. A method for improving accuracy of reduction of different skin colors is characterized by comprising the following steps:
s10, carrying out further division on four current human skin colors (yellow, white, black and brown) all over the world through big data analysis to prepare a 24-color card of the skin color representing the human skin color all over the world;
s20, shooting raw data original images of the 24-color card with skin color under different illumination of different standard light sources, and acquiring calibration data, wherein the different standard light sources are as follows: d75, D65, D50, Tl84, CWF, U30, F, A, the different illumination intensities being: 1000lux, 600lux, 400lux, 200lux, 100lux, 60lux, 30lux, 10 lux;
s30, constructing an over-determined equation, and solving color reduction matrixes under different color temperatures and illumination intensities;
s40, determining the approximate skin color of the shot target through a face recognition algorithm, and then further improving the accuracy of target skin color reduction by giving higher weight to the corresponding skin color when solving a color reduction matrix;
s50, obtaining the color temperature and the illumination of a scene according to a sensor, and then applying bilinear interpolation to obtain a corresponding color reduction matrix to act on the whole image to obtain a live effect image which is closer to the skin color of the human face.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011427655.9A CN112767257A (en) | 2020-12-07 | 2020-12-07 | Method for improving accuracy of reduction of different skin colors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011427655.9A CN112767257A (en) | 2020-12-07 | 2020-12-07 | Method for improving accuracy of reduction of different skin colors |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112767257A true CN112767257A (en) | 2021-05-07 |
Family
ID=75693443
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011427655.9A Pending CN112767257A (en) | 2020-12-07 | 2020-12-07 | Method for improving accuracy of reduction of different skin colors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112767257A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115460391A (en) * | 2022-09-13 | 2022-12-09 | 浙江大华技术股份有限公司 | Image simulation method, image simulation device, storage medium and electronic device |
| WO2022267784A1 (en) * | 2021-06-25 | 2022-12-29 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for image correction |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010074352A (en) * | 2008-09-17 | 2010-04-02 | Konica Minolta Business Technologies Inc | Color adjustment apparatus, color adjustment method, and program |
| CN104720813A (en) * | 2015-03-12 | 2015-06-24 | 西安工程大学 | Obtaining method of standard color atla for representing complexion and application of obtaining method |
| CN108377373A (en) * | 2018-05-10 | 2018-08-07 | 杭州雄迈集成电路技术有限公司 | A kind of color rendition device and method pixel-based |
| CN110751607A (en) * | 2019-10-21 | 2020-02-04 | 浙江大华技术股份有限公司 | Skin color correction method and device, storage medium and electronic device |
| CN111986151A (en) * | 2020-07-17 | 2020-11-24 | 北京瑞通科悦科技有限公司 | Skin color detection method and device |
-
2020
- 2020-12-07 CN CN202011427655.9A patent/CN112767257A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010074352A (en) * | 2008-09-17 | 2010-04-02 | Konica Minolta Business Technologies Inc | Color adjustment apparatus, color adjustment method, and program |
| CN104720813A (en) * | 2015-03-12 | 2015-06-24 | 西安工程大学 | Obtaining method of standard color atla for representing complexion and application of obtaining method |
| CN108377373A (en) * | 2018-05-10 | 2018-08-07 | 杭州雄迈集成电路技术有限公司 | A kind of color rendition device and method pixel-based |
| CN110751607A (en) * | 2019-10-21 | 2020-02-04 | 浙江大华技术股份有限公司 | Skin color correction method and device, storage medium and electronic device |
| CN111986151A (en) * | 2020-07-17 | 2020-11-24 | 北京瑞通科悦科技有限公司 | Skin color detection method and device |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022267784A1 (en) * | 2021-06-25 | 2022-12-29 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for image correction |
| CN115460391A (en) * | 2022-09-13 | 2022-12-09 | 浙江大华技术股份有限公司 | Image simulation method, image simulation device, storage medium and electronic device |
| CN115460391B (en) * | 2022-09-13 | 2024-04-16 | 浙江大华技术股份有限公司 | Image simulation method and device, storage medium and electronic device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP2227898B1 (en) | Image sensor apparatus and method for scene illuminant estimation | |
| CN104363434B (en) | Image processing apparatus | |
| RU2013114372A (en) | AUTOMATIC PROCESSING OF WHITE BALANCE WITH FLEXIBLE SELECTION OF COLOR SPACE | |
| CN105578166B (en) | A kind of colour temperature determines method and device | |
| CN112767257A (en) | Method for improving accuracy of reduction of different skin colors | |
| Kao et al. | Design considerations of color image processing pipeline for digital cameras | |
| Wang et al. | Fast automatic white balancing method by color histogram stretching | |
| JP2978615B2 (en) | Apparatus and method for adjusting color balance | |
| TW201830337A (en) | Method and device for performing automatic white balance on an image | |
| CN113556526A (en) | RGBW filter array-based color enhancement method for color night vision equipment | |
| TWI649724B (en) | Method and apparatus for determining a light source of an image and performing color vision adaptation on the image | |
| JP2005045446A (en) | Color conversion matrix calculation method and color correction method | |
| Lam et al. | Automatic white balancing using luminance component and standard deviation of RGB components [image preprocessing] | |
| Kao et al. | Color reproduction for digital imaging systems | |
| US20130329056A1 (en) | Color calibration of an image capture device in a way that is adaptive to the scene to be captured | |
| Bianco et al. | Computational color constancy | |
| JP2006222783A (en) | Preparation of color conversion table | |
| Cooper | Color segmentation as an aid to white balancing for digital still cameras | |
| CN112422940A (en) | Self-adaptive color correction method | |
| CN107454374B (en) | Image processing apparatus, image processing method, and image processing system | |
| CN115190282B (en) | Method for avoiding white balance distortion of camera module under high color temperature | |
| CN112396570B (en) | Camouflage design method | |
| JP4136820B2 (en) | Color conversion matrix calculation method and color correction method | |
| JP5983137B2 (en) | Image processing apparatus, imaging apparatus, and program | |
| JP5983138B2 (en) | Image processing apparatus, imaging apparatus, and program |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210507 |
|
| WD01 | Invention patent application deemed withdrawn after publication |