A skin color detection method and device

Abstract

This disclosure provides a skin color detection method and apparatus. It includes: establishing a human skin color chart database; acquiring the R, G, and B chromaticity values ​​of a skin image and correcting these values; converting the corrected R, G, and B chromaticity values ​​of the skin image into skin color values, and calculating the overall skin color statistical value of the skin image; and selecting the skin color number closest to the overall skin color statistical value from the human skin color chart database based on the overall skin color statistical value. This method improves the accuracy of the acquired skin image data by correcting the R, G, and B chromaticity values ​​of the acquired skin image. Furthermore, by establishing a human skin color chart database based on a standard human skin color chart, it ensures more accurate identification of the detected skin color number, facilitating more precise adjustment of the laser parameters of the cosmetic medical equipment based on the skin color number to achieve the best cosmetic effect.

Description

Technical Field

[0001] This disclosure relates to a skin color detection method and an apparatus for implementing the skin color detection method, belonging to the field of skin color detection technology. Background Technology

[0002] With the continuous development of image recognition technology, various products based on image recognition technology have emerged, which have not only brought great convenience to people's lives, but also improved their quality of life.

[0003] Solving skin color identification using image recognition technology is a pressing issue in the skincare industry. This technology needs to address problems such as accurately dividing acquired skin images into skin and background. The accuracy of this classification directly impacts subsequent processing steps, such as skin color recognition and skin tone number comparison.

[0004] Therefore, improving the accuracy of skin color detection is of significant research importance. On the other hand, accurate skin color detection is extremely difficult. For example, subtle differences in human skin color, colored lighting, shadows, strong light irradiation, and CCD color shift all affect the accurate identification of skin color. Skin color detection is particularly challenging under complex lighting conditions. Summary of the Invention

[0005] The primary technical problem to be solved by this disclosure is to provide a skin color detection method.

[0006] Another technical problem to be solved by this disclosure is to provide an apparatus for implementing the above-described skin color detection method.

[0007] To achieve the above objectives, the present disclosure adopts the following technical solution:

[0008] According to a first aspect of the present disclosure, a skin color detection method is provided, comprising the following steps:

[0009] Design a color chart matrix;

[0010] Establish a database of human skin color charts;

[0011] Obtain the R, G, and B chromaticity values ​​of the skin image and correct these chromaticity values;

[0012] After converting the R, G, and B chromaticity values ​​of the corrected skin image into skin color values, the overall skin color statistics of the skin image are calculated.

[0013] Based on the overall skin tone statistics of the skin tone image, the color number of the human skin tone that is closest to the overall skin tone statistics is selected from the human skin tone color chart database.

[0014] Preferably, the color card matrix consists of N*N code elements, where N is a positive integer. Positioning code elements are set at any three corner positions of the color card matrix. A window consisting of M*M code elements is set at the center position of the color card matrix, where M is a positive integer. Black, white, red, green, and blue reference colors are added to the color card matrix respectively.

[0015] Preferably, the human skin color chart database consists of skin color values ​​corresponding to each color number in the standard human skin color chart. The skin color value corresponding to each color number is calculated based on the R, G, and B chromaticity values ​​corresponding to each color number in the standard human skin color chart, and by the following formula.

[0016] Gray = 0.3R + 0.59G + 0.11B

[0017] Ideally, the skin tone value corresponding to each color number is multiplied by a preset number before being stored.

[0018] Preferably, the R, G, and B chromaticity values ​​of the skin image are obtained, and the chromaticity values ​​are corrected, including the following steps:

[0019] Obtain the R, G, and B chromaticity values ​​of the color chart matrix and the skin image within the window;

[0020] The color chart matrix and the skin image within the window are located to obtain the R, G, and B chromaticity values ​​of the skin image within the window;

[0021] The R, G, and B chromaticity values ​​of the skin image within the window are corrected.

[0022] Preferably, when correcting the R, G, B chromaticity values ​​of the skin image within the window, the R, G, B chromaticity values ​​of the skin image within the window are multiplied by the corresponding correction coefficients for the R, G, B chromaticity values, where the correction coefficients for the R, G, B chromaticity values ​​are the ratios of the original R, G, B chromaticity values ​​to the R, G, B chromaticity values ​​of the actually acquired color card matrix.

[0023] Preferably, when converting the R, G, and B chromaticity values ​​of the skin image within the corrected window into skin color values, sampling is performed at preset intervals to obtain the skin color values ​​corresponding to the skin image. These values ​​are then added together and averaged to obtain the overall skin color statistical value of the skin image.

[0024] Preferably, if the skin color value corresponding to the color number in the human skin color chart database is increased by a preset multiple, then the overall skin color statistical value of the skin color image is increased by the same multiple as the skin color value corresponding to the color number.

[0025] According to a second aspect of the present disclosure, a method for adjusting laser parameters of a cosmetic medical device is provided, wherein the absorbance of the detected skin color is determined using the aforementioned skin color detection method; the method for adjusting laser parameters of the cosmetic medical device further includes the following steps:

[0026] The laser parameters of the cosmetic medical device are adjusted based on the light absorption rate of the detected skin color.

[0027] According to a third aspect of the present disclosure, a skin color detection device is provided, including a color card matrix, a sealed enclosure, an image acquisition device, and a processing device. The color card matrix is ​​disposed at the detection port position of the sealed enclosure, a light source is disposed in the sealed enclosure, and the image acquisition device is connected to the processing device.

[0028] The processing device is used to establish a human skin color chart database, obtain the R, G, and B chromaticity values ​​of skin images, correct the chromaticity values, convert the corrected R, G, and B chromaticity values ​​of the skin images into skin color values, calculate the overall skin color statistical value of the skin image, and then select the human skin color number that is closest to the overall skin color statistical value from the human skin color chart database based on the overall skin color statistical value of the skin image.

[0029] The skin color detection method and apparatus disclosed herein, on the one hand, improve the accuracy of acquired skin image data by designing a color chart matrix to correct the R, G, and B chromaticity values ​​of the acquired skin images. On the other hand, a human skin color chart database is established based on a standard human skin color chart to ensure more accurate identification of the detected skin color number. This facilitates more precise adjustment of the laser parameters of the cosmetic medical equipment according to the skin color number, thereby achieving the best cosmetic effect. Attached Figure Description

[0030] Figure 1 A flowchart of the skin color detection method provided in the embodiments of this disclosure;

[0031] Figure 2 This is a schematic diagram of a color chart matrix designed in the skin color detection method provided in the embodiments of this disclosure;

[0032] Figure 3 This is a schematic diagram of the skin color detection device provided in an embodiment of the present disclosure. Detailed Implementation

[0033] The technical content of this disclosure will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] To improve the accuracy of skin color detection, such as Figure 1 As shown in the figure, this disclosure provides a skin color detection method, including the following steps:

[0035] Step S1: Design the color chart matrix.

[0036] In the embodiments of this disclosure, the color chart matrix consists of N*N code elements in two directions, where N is a positive integer. The size of the color chart matrix is ​​determined by the size of the viewport actually used to acquire the skin image. For example, as... Figure 2 As shown, when the skin image to be acquired consists of 50*50 pixels, the size of the window used to acquire the image is the same as the size of the matrix composed of 50*50 pixels. In this case, the designed color card matrix can be composed of 100*100 code elements.

[0037] In the designed color card matrix, three corner positions are arbitrarily selected to set the positioning code element 1 (e.g. Figure 2 The arbitrary selected positioning code positions shown are used for subsequent positioning of the acquired color chart matrix and the skin image within the window. A window 2 consisting of M*M code elements is set at the center of the color chart matrix, where M is a positive integer. The size of window 2 is determined by the size of the skin image to be acquired. Furthermore, black, white, red, green, and blue reference colors are added to the color chart matrix. The area of ​​each reference color added to the color chart matrix is ​​determined according to the actual acquired skin image, as long as the reference colors added to the color chart matrix are not staggered. For example, as shown... Figure 2 As shown, black, white, red, green, and blue reference colors are added to the color chart matrix in rows, and black reference colors are arranged in rectangular areas enclosed by the white, red, green, and blue reference colors. The middle area of ​​the black reference color area is left empty with M*M code elements to form window 2. The black, white, red, green, and blue reference colors added to the color chart matrix are used to correct the skin images in the subsequently acquired color chart matrix and window.

[0038] Step S2: Establish a human skin color chart database.

[0039] A human skin tone color database is established based on a standard human skin tone color chart. In the embodiments of this disclosure, the Pantone skin tone color chart is preferably used. The Pantone skin tone color chart collects 110 international standard skin tones, and its color codes are mainly represented by scientifically measuring the actual skin tone response of human skin types across the entire spectrum to indicate the closest skin color to the body. The Pantone skin tone color chart is like a comprehensive, visualized library of human skin tones, serving as a reference for skin color. This Pantone skin tone color chart is currently the only internationally available color standard that can accurately match various skin tones.

[0040] Based on the R, G, and B chromaticity values ​​corresponding to the 110 shades in the Pantone skin tone chart, the skin tone value (also known as the grayscale value) corresponding to each shade is calculated. In one embodiment of this disclosure, the skin tone value corresponding to each shade is calculated according to the following formula:

[0041] Gray = 0.3R + 0.59G + 0.11B (1)

[0042] Taking the Pantone skin tone color code 1Y02 SP as an example, the R, G, and B chromaticity values ​​(200, 169, 149) can be calculated using formula (1): Gray = 0.3 * 200 + 0.59 * 169 + 0.11 * 149 = 176.1. Similarly, formula (1) can be used to calculate the skin tone values ​​corresponding to the 110 colors in the Pantone skin tone color chart. To facilitate comparison with the acquired skin images, the skin tone values ​​corresponding to each color code can be multiplied by 100 and stored to obtain a skin tone color chart database. Table 1 shows the skin tone values ​​corresponding to some of the multiplied colors. It should be noted that the multiplication factor of the skin tone value corresponding to each color code can be set according to the requirements, as long as the skin tone value corresponding to each color code is an integer.

[0043] Table 1. Pantone Skin Tone Codes and Corresponding Skin Values

[0044] Skintone R G B grayscale value * 100 1Y01 200 172 153 17831 2Y01 203 171 153 17862 3Y01 202 171 150 17799 4Y01 203 172 149 17877 5Y01 201 173 149 17876 1Y02 200 169 149 17610 2Y02 201 169 150 17651 3Y02 200 169 148 17599 4Y02 199 169 146 17547 5Y02 197 169 145 17476 1R02 199 168 150 17532 1Y03 197 166 145 17299

[0045] Step S3: Obtain the R, G, and B chromaticity values ​​of the skin image and correct these chromaticity values.

[0046] This step is achieved through the following sub-steps:

[0047] Step S31: Obtain the R, G, and B chromaticity values ​​of the color chart matrix and the skin image within the window.

[0048] Place the window 2 of the color chart matrix designed in step S1 close to the skin, ensuring the entire color chart matrix is ​​completely illuminated by uniformly diffused white light. Then, use an image acquisition device to capture images of the color chart matrix and the skin within the window to obtain the R, G, and B chromaticity values ​​of the skin images within the color chart matrix and window. For example, as shown... Figure 2 The window 2 of the color card matrix, which consists of 100*100 code elements, is positioned close to the skin. The image acquisition device acquires the skin image within the color card matrix and the window to obtain the R, G, and B chromaticity values ​​of the 100*100 color card matrix and the skin image within the window.

[0049] Step S32: Locate the color card matrix and the skin image within the window to obtain the R, G, and B chromaticity values ​​of the skin image within the window.

[0050] The boundary of the color chart matrix is ​​obtained using the positioning code 1 in the color chart matrix to adjust its orientation, ensuring accurate placement. Then, according to the pre-designed black, white, red, green, and blue reference colors added to the color chart matrix, the R, G, and B chromaticity values ​​of these reference colors are extracted from the acquired color chart matrix and the skin image within the viewport. Based on the viewport position within the pre-designed color chart matrix, the R, G, and B chromaticity values ​​of the skin image within the viewport are obtained. Figure 2 As shown, the image boundary of the illuminated skin is obtained by using the innermost black band of the color chart matrix, thereby obtaining the R, G, and B chromaticity values ​​of 50*50 pixels within the viewport.

[0051] Step S33: Correct the R, G, and B chromaticity values ​​of the skin image within the window.

[0052] Due to the influence of colored lighting, shadows, strong light irradiation, and color shift of the image acquisition device, the R, G, and B chromaticity values ​​of the acquired color chart matrix and the skin image within the viewport exhibit deviations. Based on the original R, G, and B chromaticity values ​​of the color chart matrix, the actual acquired R, G, and B chromaticity values ​​of the color chart matrix are corrected, resulting in correction coefficients for each chromaticity value. These correction coefficients are the ratios of the original R, G, and B chromaticity values ​​to the actual acquired R, G, and B chromaticity values ​​of the color chart matrix. Specifically, the correction coefficient for the R chromaticity value is the ratio of the original R chromaticity value to the actual acquired R chromaticity value of the color chart matrix; the correction coefficient for the G chromaticity value is the ratio of the original R chromaticity value to the actual acquired G chromaticity value of the color chart matrix; and the correction coefficient for the B chromaticity value is the ratio of the original R chromaticity value to the actual acquired B chromaticity value of the color chart matrix.

[0053] Since the color chart matrix and the skin image within the window are acquired synchronously by the image acquisition device, the R, G, and B chromaticity values ​​of the skin image within the window can be corrected using the correction coefficients of the R, G, and B chromaticity values ​​obtained from the color chart matrix. Specifically, by multiplying the R, G, and B chromaticity values ​​of the skin image within the window by their respective correction coefficients, the corrected R, G, and B chromaticity values ​​of the skin image within the window can be obtained, thereby improving the accuracy of the acquired R, G, and B chromaticity values ​​of the skin image within the window.

[0054] Step S4: After converting the R, G, and B chromaticity values ​​of the corrected skin image into skin color values, calculate the overall skin color statistics of the skin image.

[0055] According to formula (1), the R, G, and B chromaticity values ​​of the skin image within the corrected window are converted into skin color values, thereby obtaining the skin color values ​​k1...k corresponding to the skin image. n(n is the number of pixels in the acquired skin image). That is, the R, G, and B chromaticity values ​​of the corrected 50*50 pixel points are converted into corresponding skin color values.

[0056] To improve the conversion efficiency of R, G, B chromaticity values ​​of corrected skin images into skin color values, sampling can be performed at preset intervals. For example, the R, G, B chromaticity values ​​of a corrected 50*50 pixel image can be converted into skin color values ​​every 25 pixels, resulting in a matrix of skin color values ​​for a 10*10 pixel image.

[0057] The skin color values ​​k1...k corresponding to the skin image n The sums are then averaged to obtain the overall skin tone statistical value of the skin tone image. Specifically, the overall skin tone statistical value of the skin tone image is obtained according to the following formula (2). If the skin tone value corresponding to the color number in the human skin tone color chart database is increased by a preset multiple, then the overall skin tone statistical value of the skin tone image is increased by the same multiple as the skin tone value corresponding to the color number.

[0058]

[0059] Step S5: Based on the overall skin tone statistics of the skin tone image, select the skin tone color number from the human skin tone color chart database that is closest to the overall skin tone statistics.

[0060] The overall skin color statistics of the skin color image obtained in step S4 are compared with the skin color values ​​corresponding to the color numbers in the human skin color chart database. The color number that is closest to the overall skin color value is selected as the color number of the detected skin.

[0061] In another embodiment of this disclosure, a method for adjusting laser parameters of a cosmetic medical device is provided, employing the above-mentioned... Figure 1 and Figure 2 The corresponding skin color detection method determines the absorbance of the detected skin color; the method for adjusting the laser parameters of this cosmetic medical device also includes:

[0062] Adjust the laser parameters of the cosmetic medical equipment based on the light absorption rate of the detected skin color.

[0063] Specifically, the absorbance of skin color is categorized into strong, weak, and medium levels.

[0064] Adjusting the laser parameters of cosmetic medical equipment based on the absorbance of skin color can include:

[0065] When the light absorption rate of skin tone is high, adjust the laser parameters of the cosmetic medical equipment to reduce the output laser energy.

[0066] When the absorbance of skin color is weak, adjust the laser parameters of the cosmetic medical equipment to increase the output laser energy.

[0067] When the absorbance of skin tone is moderate, maintain the laser parameters of the cosmetic medical equipment and output laser energy normally.

[0068] like Figure 3 As shown, this embodiment of the present disclosure also provides a skin color detection device, including a color chart matrix 10, an opaque sealed cover 20, an image acquisition device 30, and a processing device 40. The color chart matrix 10 is disposed at the detection port position of the opaque sealed cover 20, a light source 50 is disposed in the sealed cover, and the image acquisition device 30 is connected to the processing device 40.

[0069] The color card matrix 10 consists of N*N code elements in two directions (horizontal and vertical), where N is a positive integer. Positioning code element 1 is set at any of the three corner positions of the color card matrix 10 (e.g., ...). Figure 2 The arbitrary selected positioning code position shown is used for subsequent positioning of the acquired color chart matrix and the skin image within the window. A window 2 consisting of M*M code elements, where M is a positive integer, is set at the center of the color chart matrix 10. Furthermore, black, white, red, green, and blue reference colors are added to the color chart matrix 10. The area of ​​each reference color added to the color chart matrix 10 depends on the actual acquired skin image, as long as the reference colors added to the color chart matrix 10 are not staggered.

[0070] The light source set in the opaque sealed enclosure 20 can be an LED light source, which is used to ensure that the color card matrix 10 is completely in a uniformly diffuse white light illumination environment.

[0071] The image acquisition device 30 can be implemented using a CCD camera or a CMOS camera. This image acquisition device 30 is used to acquire skin images within the color chart matrix and the viewing window, and output the R, G, and B chromaticity values ​​of the skin images within the color chart matrix and the viewing window to the processing device 40.

[0072] Specifically, the window 2 of the color chart matrix 10 of this skin color detection device is placed close to the skin, and the skin image within the color chart matrix and the window is acquired by the image acquisition device to obtain the R, G, and B chromaticity values ​​of the skin image within the color chart matrix and the window.

[0073] The processing device 40 can be implemented using an industrial control computer or a microcontroller. This processing device 40 is used to establish a human skin color chart database, acquire the R, G, and B chromaticity values ​​of the skin image, correct these chromaticity values, convert the corrected R, G, and B chromaticity values ​​of the skin image into skin color values, calculate the overall skin color statistical value of the skin image, and then select the closest human skin color number from the human skin color chart database based on the overall skin color statistical value of the skin image. The process of processing the acquired R, G, and B chromaticity values ​​of the skin image through the processing device to finally confirm the skin color number of the detected skin is the same as steps S3 to S5, and will not be described again here.

[0074] The skin color detection method and apparatus disclosed herein, on the one hand, improve the accuracy of acquired skin image data by designing a color chart matrix to correct the R, G, and B chromaticity values ​​of the acquired skin images. On the other hand, a human skin color chart database is established based on the Pantone human skin color chart to ensure more accurate identification of the detected skin color number. This facilitates more precise adjustment of the laser parameters of the cosmetic medical equipment according to the skin color number, thereby achieving the best cosmetic effect.

[0075] The skin color detection method and apparatus provided in this disclosure have been described in detail above. Any obvious modifications made to this disclosure by those skilled in the art without departing from its essential content will fall within the scope of protection of this patent.

Claims

1. A skin color detection method, characterized in that... Includes the following steps: Design a color chart matrix; Establish a database of human skin color charts; Obtain the R, G, and B chromaticity values ​​of the skin image and correct these chromaticity values; After converting the R, G, and B chromaticity values ​​of the corrected skin image into skin color values, the overall skin color statistics of the skin image are calculated. Based on the overall skin tone statistics of the skin tone image, the color number of the human skin tone that is closest to the overall skin tone statistics is selected from the human skin tone color card database, so as to adjust the laser parameters of the cosmetic medical device according to the color number of the skin tone. The color card matrix consists of N*N code elements, where N is a positive integer. Three corner positions of the color card matrix are arbitrarily selected to set positioning code elements. A window consisting of M*M code elements is set at the center position of the color card matrix, where M is a positive integer. Black, white, red, green and blue reference colors are added to the color card matrix respectively. The step of obtaining the R, G, and B chromaticity values ​​of the skin image and correcting these chromaticity values ​​includes the following steps: Obtain the R, G, and B chromaticity values ​​of the color chart matrix and the skin image within the window; The color chart matrix and the skin image within the window are located to obtain the R, G, and B chromaticity values ​​of the skin image within the window; The R, G, and B chromaticity values ​​of the skin image within the window are corrected; The human skin tone color chart database consists of skin tone values ​​corresponding to each color number in the standard human skin tone color chart. The skin tone value corresponding to each color number is calculated based on the R, G, and B chromaticity values ​​corresponding to each color number in the standard human skin tone color chart, and using the following formula: Gray = 0.3R + 0.59G + 0.11B Each skin tone value is multiplied by a preset factor and then stored. When correcting the R, G, B chromaticity values ​​of the skin image within the window, the R, G, B chromaticity values ​​of the skin image within the window are multiplied by the corresponding correction coefficients for the R, G, B chromaticity values. The correction coefficients for the R, G, B chromaticity values ​​are the ratios of the original R, G, B chromaticity values ​​to the R, G, B chromaticity values ​​of the actually acquired color card matrix.

2. The skin color detection method as described in claim 1, characterized in that: When converting the R, G, and B chromaticity values ​​of the skin image within the corrected window into skin color values, sampling is performed at preset intervals. The skin color values ​​corresponding to the skin image are then added together and averaged to obtain the overall skin color statistical value of the skin image.

3. The skin color detection method as described in claim 1, characterized in that: When acquiring the R, G, and B chromaticity values ​​of the skin image, the color chart matrix is ​​completely under uniform diffuse white light illumination.

4. The skin color detection method as described in claim 1, characterized in that: If the skin color value corresponding to the color number in the human skin color chart database is increased by a preset multiple, then the overall skin color statistical value of the skin color image is increased by the same multiple as the skin color value corresponding to the color number.

5. The skin color detection method as described in claim 4, characterized in that: The color chart matrix consists of a rectangular area enclosed by white, red, green, and blue reference colors, with black reference colors arranged in it. The middle area of ​​the black reference color area is left empty with M*M code elements to form a window.

6. A method for adjusting laser parameters of a cosmetic medical device, characterized in that... The absorbance of the detected skin color is determined using the skin color detection method described in any one of claims 1 to 5; the method for adjusting the laser parameters of the cosmetic medical device further includes: The laser parameters of the cosmetic medical device are adjusted based on the light absorption rate of the detected skin color.

7. A skin color detection device, characterized in that... It includes a color card matrix, a sealed enclosure, an image acquisition device, and a processing device. The color card matrix is ​​located at the detection port of the sealed enclosure, a light source is provided in the sealed enclosure, and the image acquisition device is connected to the processing device. The processing device is used to establish a human skin color chart database, obtain the R, G, and B chromaticity values ​​of skin images, correct the chromaticity values, convert the corrected R, G, and B chromaticity values ​​of the skin images into skin color values, calculate the overall skin color statistical value of the skin color image, and then select the human skin color number that is closest to the overall skin color statistical value from the human skin color chart database based on the overall skin color statistical value, so as to adjust the laser parameters of the medical device according to the skin color number. The color card matrix consists of N*N code elements, where N is a positive integer. Three corner positions of the color card matrix are arbitrarily selected to set positioning code elements. A window consisting of M*M code elements is set at the center position of the color card matrix, where M is a positive integer. Black, white, red, green and blue reference colors are added to the color card matrix respectively. The step of obtaining the R, G, and B chromaticity values ​​of the skin image and correcting these chromaticity values ​​includes the following steps: Obtain the R, G, and B chromaticity values ​​of the color chart matrix and the skin image within the viewport; The color chart matrix and the skin image within the window are located to obtain the R, G, and B chromaticity values ​​of the skin image within the window; The R, G, and B chromaticity values ​​of the skin image within the window are corrected; The human skin tone color chart database consists of skin tone values ​​corresponding to each color number in the standard human skin tone color chart. The skin tone value corresponding to each color number is calculated based on the R, G, and B chromaticity values ​​corresponding to each color number in the standard human skin tone color chart, and using the following formula: Gray = 0.3R + 0.59G + 0.11B Each skin tone value is multiplied by a preset factor and then stored. When correcting the R, G, B chromaticity values ​​of the skin image within the window, the R, G, B chromaticity values ​​of the skin image within the window are multiplied by the corresponding correction coefficients for the R, G, B chromaticity values. The correction coefficients for the R, G, B chromaticity values ​​are the ratios of the original R, G, B chromaticity values ​​to the R, G, B chromaticity values ​​of the actually acquired color card matrix.

Images