An automatic sensitive skin discrimination method combining skin color and acne distribution

Abstract

The application belongs to the technical field of skin detection, and particularly relates to a sensitive skin automatic identification method combining skin color and acne distribution, which comprises the following steps: collecting a user's face image through a mobile terminal; identifying a skin color type based on a LAB color space; extracting red region values and brightness values of each region of the face by using an HSV color model; positioning a face acne region based on a lightweight target detection network; respectively calculating acne densities of each region of the face; calculating a region sensitivity index based on the red region values, the acne densities and the skin color type of each region; synchronously outputting red region values and skin types of each region of the user's face by collecting the user's face image and using a double color space parallel processing mode, and obtaining a sensitivity index capable of reflecting the sensitivity degree of each region of the user's face based on the calculation of the acne densities of each region of the face, wherein the skin quality correction coefficient corresponding to the skin type is used to correct the sensitivity index, which can reduce missed diagnosis and misdiagnosis caused by skin color differences to a certain extent.

Description

Technical Field

[0001] This invention belongs to the field of skin detection technology, specifically a method for automatically identifying sensitive skin by combining skin color and acne distribution. Background Technology

[0002] The accurate evaluation of facial skin quality plays a crucial role in the beauty industry, dermatology, and plastic surgery. Currently, skin testing instruments on the market, both domestically and internationally, suffer from drawbacks such as high cost, bulky size, and inconvenience. Manual evaluation is difficult to achieve consistently, as different individuals typically provide varying assessments of the same skin sample. For dermatologists, eliminating subjective bias and providing objective results is paramount.

[0003] In existing technologies, facial images of users are collected via mobile terminals, and then image processing is used to analyze the condition of sensitive facial skin. Based on this, users can determine whether their face is sensitive, the specific sensitive areas, and receive appropriate care based on the sensitivity level. The skin condition after care is continuously tracked, thus achieving full-process skin management from detection to care to feedback. However, existing technologies rely solely on single image processing methods, such as converting the image only to the HSV color space to obtain the red zone value corresponding to the user's facial image, and then analyzing the sensitivity index based on the red zone value. This process ignores the influence of skin texture on image processing, which may result in the output sensitivity index not accurately reflecting the user's true skin sensitivity state, leading to misdiagnosis and missed diagnosis.

[0004] Therefore, the present invention provides an automatic method for identifying sensitive skin by combining skin color and acne distribution. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for automatically identifying sensitive skin by combining skin color and acne distribution, comprising the following steps:

[0007] S1: Collect user facial images via mobile terminal;

[0008] S2: Skin tone type identification based on LAB color space;

[0009] S3: Use the HSV color model to extract the red zone values ​​of each area of ​​the face. and brightness value ;

[0010] S4: Localize facial acne areas using a lightweight object detection network;

[0011] S5: Calculate the acne density in each area of ​​the face. ;

[0012] S6: Red zone values ​​based on each region Acne density And skin color type to calculate the area sensitivity index ;

[0013] S7: Based on the sensitivity index of each region With set threshold The comparison results are used to output the sensitivity determination result.

[0014] Preferably, the method for identifying skin color type is as follows:

[0015] Calculate lightness in the LAB color space Values ​​and the yellow and blue axes The ratio of values;

[0016] According to skin color classification rules:

[0017] type ;

[0018] according to Determine the user's skin tone type.

[0019] Preferably, in the HSV color model conversion:

[0020] Red zone value Calculated by the percentage of pixels in the H channel in the 0-10° and 170-180° ranges;

[0021] Brightness value Used to eliminate light interference, when Start the supplemental lighting at the appropriate time;

[0022] The facial region includes the face area, the T-zone, and the contour area.

[0023] Preferably, the pimple density The calculations include:

[0024] Each region uses an independent YOLOv5s model for detection:

[0025] The facial area: input size 416×416, confidence level 0.7;

[0026] The T region has an input size of 320×320 and a confidence level of 0.6.

[0027] The contour region has the following input dimensions: 256×256, and a confidence level of 0.5.

[0028] According to the formula:

[0029] ；

[0030] in, For the region The number of pimples; For the region The area.

[0031] Preferably, the method for calculating the regional sensitivity index is as follows:

[0032] Extracting red zone values ​​from various areas of the face and acne density ;

[0033] According to the formula:

[0034] ;

[0035] in The weighting coefficient for the red zone value. This is the acne density weighting coefficient. This is the skin tone correction factor; where, .

[0036] Preferably, the method for outputting the sensitivity determination result is as follows:

[0037] The calculated regional sensitivity index With set threshold Comparison:

[0038] when At that time, the skin in the affected area is determined to be sensitive, and the number of sensitive areas is... If a certain area, such as the cheek, is sensitive, then the entire face is considered sensitive.

[0039] Among them, according to the regional sensitivity index The size of the value determines the sensitivity level of the corresponding region, including:

[0040] like It was determined to be healthy skin in the region;

[0041] like It was determined to be moderately sensitive in the region;

[0042] like The area was determined to be highly sensitive.

[0043] Then, based on the regional sensitivity level, the facial sensitivity level is output, including:

[0044] If any area is severely sensitive, then the face is severely sensitive.

[0045] Otherwise, if any area is moderately sensitive, then the face is moderately sensitive;

[0046] Otherwise, it's healthy skin.

[0047] Preferably, the set threshold Basic threshold The result after dynamic correction of the user's skin type;

[0048] The user's skin type is based on moisture level. To make distinctions, including:

[0049] Skin type ;

[0050] in, Represents any area of ​​the user's face Moisture content;

[0051] For dry regions, ;

[0052] For dry regions, then ;

[0053] For the neutral region, then .

[0054] Preferred options also include:

[0055] S8: Based on the judgment result and sensitivity level, call the preset plan library; the preset plan library stores nursing plans corresponding to the sensitivity level, including:

[0056] Identify the user's facial sensitivity level and retrieve the corresponding care plan.

[0057] Preferred options also include:

[0058] S9: Establish a user profile database to record the user's initial sensitive skin assessment results and facial sensitivity level;

[0059] The system tracks and detects users at equal time intervals, and outputs real-time judgment results and facial sensitivity levels.

[0060] Based on the real-time judgment results and facial sensitivity level, a change curve is output.

[0061] The beneficial effects of this invention are as follows:

[0062] 1. The present invention provides an automatic sensitive skin identification method that combines skin color and acne distribution. By acquiring user facial images and using a dual-color space parallel processing mode, the red zone value and skin type of each area of ​​the user's face are output simultaneously. Based on the calculation of acne density in each area of ​​the face, a sensitivity index that reflects the sensitivity of each area of ​​the user's face is obtained. The sensitivity index is corrected by a skin type correction coefficient corresponding to the skin type, which can reduce the missed diagnosis and misdiagnosis caused by skin color differences to a certain extent. Attached Figure Description

[0063] The invention will now be further described with reference to the accompanying drawings.

[0064] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0065] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0066] like Figure 1 As shown in the embodiment of the present invention, an automatic sensitive skin identification method combining skin color and acne distribution includes the following steps:

[0067] S1: Collect user facial images via mobile terminal;

[0068] S2: Skin tone type identification based on LAB color space;

[0069] S3: Use the HSV color model to extract the red zone values ​​of each area of ​​the face. and brightness value ;

[0070] S4: Localize facial acne areas using a lightweight object detection network;

[0071] S5: Calculate the acne density in each area of ​​the face. ;

[0072] S6: Red zone values ​​based on each region Acne density And skin color type to calculate the area sensitivity index ;

[0073] S7: Based on the sensitivity index of each region With set threshold The comparison results are used to output the sensitivity determination result.

[0074] In existing technologies, only a single image processing method is relied upon, such as converting the image to the HSV color space to obtain the red zone value corresponding to the user's facial image, and then analyzing the sensitivity index based on the red zone value. However, in the above process, the influence of skin texture on image processing is ignored, which may result in the output sensitivity index not being able to truly reflect the user's real skin sensitivity state, thus causing misdiagnosis and missed diagnosis.

[0075] In one embodiment of the present invention, when a user performs sensitive skin detection, facial images can be captured via a mobile terminal. Specifically, in one embodiment, the mobile terminal is a mobile phone. The user opens the APP to start the detection, the front camera automatically turns on, and captures a high-definition image of the user's face with a resolution of at least 1080P. The system performs parallel processing in two color spaces, including HSV space processing and LAB space processing, to extract the red zone values ​​of each region of the face. and brightness value and calculation The system uses a lightweight object detection network to locate acne-prone areas on the face, and calculates the acne density in each area according to a specific formula, based on the values ​​and skin tone type. Finally, based on the red zone values ​​of each region... With pimple density Calculate the regional sensitivity index and according to Values ​​corrected for sensitivity indices in each region Sensitivity indices for each region With set threshold The comparison results can then output the sensitive skin assessment result. It's understandable that in traditional sensitive skin detection, the false negative rate is higher in darker-skinned individuals compared to lighter-skinned individuals. This is because the skin of darker-skinned individuals is affected by melanin, leading to interference with the red area signal, thus requiring compensation processing. Conversely, lighter-skinned individuals have a thinner dermis, resulting in a stronger red area signal, necessitating enhanced detection. Based on this, by performing parallel processing in a dual color space, the red area values ​​of different regions on the user's face are identified. and brightness value In addition, it is also based on The system calculates the sensitivity index of different areas of the user's face, identifies the user's skin tone, and then calculates the sensitivity index of each area. Based on facial skin tone type, correct the sensitivity index of each area. This reduces the interference of red zone signals caused by melanin interference, thus improving the accuracy of sensitive skin detection for people with darker skin tones.

[0076] In one embodiment, the method for identifying skin color type is as follows:

[0077] Calculating lightness in the LAB color space Values ​​and Yellow / Blue Axis The ratio of values;

[0078] According to skin color classification rules:

[0079] type ;

[0080] according to Determine the user's skin tone type.

[0081] After calculating the user's skin tone type using the LAB color space, a corresponding skin tone correction coefficient is matched to the user based on the identified skin tone type. Adjust the coefficient according to the corresponding skin tone. This can correct for missed diagnoses caused by skin color during subsequent sensitivity index calculations for different areas of the user's face. In one embodiment, the skin color correction coefficient... The selection rules are as follows:

[0082] For type - , ;

[0083] For type - , ;

[0084] For type - , ;

[0085] In one embodiment, during the HSV color model conversion:

[0086] Red zone value Calculated by the percentage of pixels in the H channel in the 0-10° and 170-180° ranges;

[0087] Brightness value Used to eliminate light interference, when Start the supplemental lighting at the appropriate time;

[0088] The facial region includes the face area, the T-zone, and the contour area.

[0089] In one embodiment, the pimple density The calculations include:

[0090] Each region uses an independent YOLOv5s model for detection:

[0091] The facial area: input size 416×416, confidence level 0.7;

[0092] The T region has an input size of 320×320 and a confidence level of 0.6.

[0093] The contour region has the following input dimensions: 256×256, and a confidence level of 0.5.

[0094] According to the formula:

[0095] ；

[0096] in, For the region The number of pimples; For the region The area.

[0097] Based on the above, in order to quantify the sensitivity index of different areas of a user's face, it is necessary to measure the acne density in each area of ​​the user's face separately. The calculations will be explained below using specific data:

[0098] Assume the user's skin data is as follows:

[0099] Skin tone type V;

[0100] Facial area: Moisture content 28. 0.85 0.40;

[0101] T zone: Moisture value 65. 0.60 0.80;

[0102] Contour area: Moisture value 45. 0.30 0.20;

[0103] Based on the above, data from various regions of the user's face are identified and recorded, including skin tone type and moisture levels in each region. And the calculated acne density in each region. .

[0104] In one embodiment, the method for calculating the regional sensitivity index is as follows:

[0105] Extracting red zone values ​​from various areas of the face and acne density ;

[0106] According to the formula:

[0107] ;

[0108] in The weighting coefficient for the red zone value. This is the acne density weighting coefficient. This is the skin tone correction factor; where, .

[0109] In one embodiment of the present invention, a specific scenario will be used for explanation and illustration:

[0110] Suppose there are users A and B, where:

[0111] User A is of West African descent, with a V-shaped skin tone. ;

[0112] User B is a Caucasian from Northern Europe, with a skin type I. ;

[0113] First, extract basic parameters for user A and user B:

[0114] The following parameters were obtained:

[0115] For the facial area:

[0116] User A: 0.85 0.40;

[0117] User B: 0.78 0.35;

[0118] Introducing a skin tone correction factor:

[0119] User A: ;

[0120] User B: ;

[0121] The sensitivity index calculations are compared as follows:

[0122] In the traditional approach:

[0123] ;

[0124] ;

[0125] After skin tone correction:

[0126] ;

[0127] ;

[0128] According to the sensitivity index calculated by the traditional method, the sensitivity indices of users A and B are similar. However, after skin color correction, the sensitivity indices of users A and B differ significantly. In other words, if the threshold is 0.8, according to the traditional method, neither user A nor user B has sensitive skin. However, after skin color correction, user B has sensitive skin, while user A has healthy skin. Skin color correction can reduce misjudgments and missed diagnoses caused by skin color.

[0129] Furthermore, continuing with the facial region data of user A obtained in the previous embodiment, as follows:

[0130] Skin tone type V;

[0131] Facial area: Moisture content 28. 0.85 0.40;

[0132] T zone: Moisture value 65. 0.60 0.80;

[0133] Contour area: Moisture value 45. 0.30 0.20;

[0134] The sensitivity indices of various facial regions for user A were calculated as follows:

[0135] ;

[0136] ;

[0137] ;

[0138] Based on the above calculations, the sensitivity index for each region of user A's face is obtained. Then, based on the obtained sensitivity index and a set threshold... By comparing the data, it can be determined whether different areas of a user's face are sensitive, and the facial sensitivity level of user A can be determined based on the magnitude of the sensitivity index and the number of sensitive areas.

[0139] In one embodiment, the method for outputting the sensitivity determination result is as follows:

[0140] The calculated regional sensitivity index With set threshold Comparison:

[0141] when At that time, the skin in the affected area is determined to be sensitive, and the number of sensitive areas is... If a certain area, such as the cheek, is sensitive, then the entire face is considered sensitive.

[0142] Among them, according to the regional sensitivity index The size of the value determines the sensitivity level of the corresponding region, including:

[0143] like It was determined to be healthy skin in the region;

[0144] like It was determined to be moderately sensitive in the region;

[0145] like The area was determined to be highly sensitive.

[0146] Then, based on the regional sensitivity level, the facial sensitivity level is output, including:

[0147] If any area is severely sensitive, then the face is severely sensitive.

[0148] Otherwise, if any area is moderately sensitive, then the face is moderately sensitive;

[0149] Otherwise, it's healthy skin.

[0150] As mentioned above, after calculating the sensitivity index of each area of ​​the user's face, it is also necessary to set a threshold. The system determines the sensitivity of different areas of the user's face. The following explanation continues using the facial data of user A from the above embodiment:

[0151] From the above, we can see that the sensitivity index of different areas of User A's face is as follows:

[0152] ;

[0153] ;

[0154] ;

[0155] Assuming a threshold is set If the sensitivity index of all areas of user A's face is lower than the set threshold, then... Therefore, it can be determined that all areas of user A's face are healthy skin. When outputting the facial sensitivity level according to the area sensitivity level, since no area of ​​user A's face is sensitive, user A's face is healthy skin.

[0156] However, according to the above embodiment, user B's cheek sensitivity index is:

[0157] ;

[0158] Therefore, it can be determined that user B's cheek is moderately sensitive. If the sensitivity indices of user B's T-zone and contour area are both below a set threshold... If so, it can be determined that user B's face is moderately sensitive, and based on this, it can be output that user A's face is healthy skin, while user B's face is moderately sensitive.

[0159] In one embodiment, the set threshold Basic threshold The result after dynamic correction of the user's skin type;

[0160] The user's skin type is based on moisture level. To make distinctions, including:

[0161] Skin type ;

[0162] in, Represents any area of ​​the user's face Moisture content;

[0163] For dry regions, then ;

[0164] For dry regions, then ;

[0165] For the neutral region, then .

[0166] In one embodiment, when a user performs a facial skin sensitivity test using a mobile app, skin moisture values ​​are obtained based on spectral analysis. Continuing with the above embodiment, user B obtained the following data through testing:

[0167] Cheek area: Based on the skin type classification rules, user B's skin type, specifically the cheek skin type, is dry.

[0168] Assuming a base threshold ;

[0169] Therefore, according to the calculation, for user B, the set threshold corresponding to their cheek area is... ;

[0170] Based on the above, the sensitivity index of user B's cheek area obtained from the detection can be determined as follows:

[0171] ;

[0172] And based on a set threshold The calculation can output a judgment result, namely that user B's cheek area is moderately sensitive.

[0173] In one embodiment, it further includes:

[0174] S8: Based on the judgment result and sensitivity level, call the preset plan library; the preset plan library stores nursing plans corresponding to the sensitivity level, including:

[0175] Identify the user's facial sensitivity level and retrieve the corresponding care plan.

[0176] Based on the above, once user B's face is identified as moderately sensitive, the system retrieves relevant skincare solutions for moderately sensitive faces from the pre-set solution library based on the judgment result and sensitivity level. For example, if user B's main sensitive area is the cheek, the corresponding product could be a Centella Asiatica soothing mask. As for the T-zone and facial contour area, since these two areas are not sensitive, basic skincare is more recommended.

[0177] In one embodiment, it further includes:

[0178] S9: Establish a user profile database to record the user's initial sensitive skin assessment results and facial sensitivity level;

[0179] The system tracks and detects users at equal time intervals, and outputs real-time judgment results and facial sensitivity levels.

[0180] Based on the real-time judgment results and facial sensitivity level, a change curve is output.

[0181] In addition to providing users with the basic function of detecting facial sensitivity levels, in one embodiment of the present invention, based on the facial sensitivity level obtained after the user's detection, long-term tracking and early warning of the user's facial sensitivity level can also be achieved. In one embodiment, if user B's initial detection shows moderate sensitivity in the cheek area, with a 30-day cycle, user B is proactively reminded to conduct another facial sensitivity test after 30 days. If the second test shows that user B's facial sensitivity index has decreased, it indicates that the care plan pushed to the user is effective. Then, the user's facial sensitivity changes are tracked, and the sensitivity index change curve of each facial area is output according to the detection time node. If user B's facial sensitivity index increases, and the sensitivity index of user B's facial area does not decrease in the next cycle, then the user needs to seek medical attention in a timely manner based on the change in the sensitivity index.

[0182] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for automatically identifying sensitive skin by combining skin tone and acne distribution, characterized in that: Includes the following steps: S1: Collect user facial images via mobile terminal; S2: Skin tone type identification based on LAB color space; S3: Use the HSV color model to extract the red zone values ​​of each area of ​​the face. and brightness value ; S4: Localize facial acne areas using a lightweight object detection network; S5: Calculate the acne density in each area of ​​the face. ; S6: Red zone values ​​based on each region Acne density And skin color type to calculate the area sensitivity index ; S7: Based on the sensitivity index of each region With set threshold The comparison results are used to output the sensitivity determination result; The method for identifying skin color type is as follows: Calculating lightness in the LAB color space Values ​​and the yellow and blue axes The ratio of values; According to skin color classification rules: type ; according to Determine the user's skin tone type; In the HSV color model conversion: Red zone value Calculated by the percentage of pixels in the H channel in the 0-10° and 170-180° ranges; Brightness value Used to eliminate light interference, when Start the supplemental lighting at the appropriate time; The facial region includes the facial area, the T-zone, and the contour area; The density of pimples The calculations include: Each region uses an independent YOLOv5s model for detection: The facial area: input size 416×416, confidence level 0.7; The T region has an input size of 320×320 and a confidence level of 0.

6. The contour region has the following input dimensions: 256×256, and a confidence level of 0.

5. According to the formula: ； in, For the region The number of pimples; For the region The area; The method for calculating the regional sensitivity index is as follows: Extracting red zone values ​​from various areas of the face and acne density ; According to the formula: ； in The weighting coefficient for the red zone value. This is the acne density weighting coefficient. This is the skin tone correction factor; where, .

2. The method for automatically identifying sensitive skin by combining skin tone and acne distribution according to claim 1, characterized in that: The method for outputting the sensitivity determination result is as follows: The calculated regional sensitivity index With set threshold Comparison: when At that time, the skin in the affected area is determined to be sensitive, and the number of sensitive areas is... If a certain area, such as the cheek, is sensitive, then the entire face is considered sensitive. Among them, according to the regional sensitivity index The size of the value determines the sensitivity level of the corresponding region, including: like It was determined to be healthy skin in the region; like It was determined to be moderately sensitive in the region; like The area was determined to be highly sensitive. Then, based on the regional sensitivity level, the facial sensitivity level is output, including: If any area is severely sensitive, then the face is severely sensitive. Otherwise, if any area is moderately sensitive, then the face is moderately sensitive; Otherwise, it's healthy skin.

3. The method for automatically identifying sensitive skin by combining skin tone and acne distribution according to claim 2, characterized in that: The set threshold Basic threshold The result after dynamic correction of the user's skin type; The user's skin type is based on moisture level. To make distinctions, including: Skin type ; in, Represents any area of ​​the user's face Moisture content; For dry regions, then ; For dry regions, then ; For the neutral region, then .

4. The method for automatically identifying sensitive skin by combining skin tone and acne distribution according to claim 3, characterized in that: Also includes: S8: Based on the judgment result and sensitivity level, call the preset solution library; The pre-set protocol library stores care protocols corresponding to sensitivity levels, including: Identify the user's facial sensitivity level and retrieve the corresponding care plan.

5. The method for automatically identifying sensitive skin by combining skin tone and acne distribution according to claim 4, characterized in that: Also includes: S9: Establish a user profile database to record the user's initial sensitive skin assessment results and facial sensitivity level; The system tracks and detects users at equal time intervals, and outputs real-time judgment results and facial sensitivity levels. Based on the real-time judgment results and facial sensitivity level, a change curve is output.

Images