A Smart Beauty Method Based on Facial Partitioning

Abstract

This invention discloses a smart beauty method based on facial segmentation, belonging to the field of facial smart beauty technology. It utilizes an image acquisition module to acquire images of user facial segments, an image processing module to calculate and output adjusted skin brightness values ​​(LF), and a skin simulation module to simulate and extrapolate the optimized skin state, summarizing the noise pixel count (XS). noise The image processing module is used to calculate and output the optimized skin smoothness PF and skin tone uniformity JF, respectively, and to extract the skin tone uniformity threshold JF corresponding to the patient's age group from the database. HZ The method is compared with the skin tone uniformity value JF, and the beauty method is output by the interactive module. This invention improves the processing accuracy and efficiency of intelligent beauty methods through precise facial partitioning, scientific parameter adjustment mechanism, efficient iterative processing process and significant improvement of skin tone uniformity. It also provides users with more personalized and accurate beauty solutions.

Description

Technical Field

[0001] This invention relates to the field of facial intelligent beauty technology, specifically to an intelligent beauty method based on facial partitioning. Background Technology

[0002] With the continuous development of image processing technology, intelligent beauty methods are increasingly widely used in the beauty industry. Among them, facial zoning is an important concept in the fields of cosmetic medicine and skin care. According to the theory of facial aesthetic zoning, the human face can be subdivided into multiple parts, such as the forehead, eyebrows, nose, lips, chin, nasolabial folds, eye sockets, cheeks, and temples. The skin characteristics, aging manifestations, and required care methods of these parts are different. Therefore, intelligent beauty methods based on facial zoning can provide precise care for different parts and meet personalized beauty needs.

[0003] However, some existing smart beauty methods often adopt a global processing strategy without personalized processing for different areas of the face, resulting in less refined results. In traditional beauty methods, parameter adjustments often rely on human experience, lacking scientific basis and automated adjustment mechanisms. Furthermore, when performing multiple iterations, existing methods often struggle to quickly converge to the optimal solution, leading to low iteration efficiency. In addition, in terms of skin tone evenness, existing methods often fail to accurately identify and address facial pigmentation and pore issues, resulting in poor skin tone evenness improvement. Summary of the Invention

[0004] The purpose of this invention is to provide a smart beauty method based on facial partitioning, which solves the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution, and the specific implementation steps are as follows:

[0006] Step 1: Use the image acquisition module to acquire images of the user's facial regions and transmit them to the image processing module and store them in the database;

[0007] Step 2: Using the image processing module, calculate and output the adjusted skin brightness value LF;

[0008] Step 3: Based on the adjusted skin brightness value LF and the custom brightness enhancement value TL, and using the skin simulation module, simulate and extrapolate the optimized skin state, and summarize the noise pixel count XS. noise ;

[0009] Step 4: Based on the number of noisy pixels XS noise The adjusted skin brightness value LF is calculated, and the optimized skin smoothness PF and skin color uniformity value JF are calculated and output using the image processing module, respectively.

[0010] Step 5: Extract the skin tone uniformity threshold JF corresponding to the patient's age group from the database. HZ The skin tone uniformity value (JF) is compared with the skin tone uniformity value (JF), and the beauty method is output by the interactive module.

[0011] The image processing module includes a unit for enhancing the brightness of facial skin regions, a unit for optimizing the smoothness of facial skin regions, and a unit for enhancing the uniformity of facial skin regions.

[0012] Preferably, the equipment used in the image acquisition module includes a high-resolution camera and professional photography equipment;

[0013] The image processing module uses devices including a computer and image processing software.

[0014] The equipment used in the skin simulation module includes a skin simulation system;

[0015] The interactive module uses devices including display devices.

[0016] Preferably, the calculation formula for enhancing the facial skin brightness unit is as follows:

[0017] LF=SQRT(R avg +G avg +B avg )×(1+TL / 100)-Y / 2;

[0018] Y = YM / FM;

[0019] in:

[0020] LF represents the adjusted skin brightness value;

[0021] SQRT stands for the radical sign;

[0022] R avg The average skin tone;

[0023] G avg Average skin radiance;

[0024] B avg Average skin transparency;

[0025] TL is a custom brightness boost value. TL reflects the degree to which the user-defined brightness needs to be boosted, and its value ranges from {0-100}.

[0026] Y represents the skin shading intensity;

[0027] YM represents the shadow area, and FM represents the total area of ​​the facial regions.

[0028] Preferably, the average skin rosiness Ravg The calculation formula is as follows:

[0029] R avg =(R1+R2+R3+......+R N ) / N;

[0030] N is the total number of pixels, which reflects the total number of pixels used for skin detection within the facial area;

[0031] R1 represents the skin tone of the first pixel, R2 represents the skin tone of the second pixel, R3 represents the skin tone of the third pixel, and R... N The skin rosiness of the Nth pixel;

[0032] The average skin luster G avg The calculation formula is as follows:

[0033] G avg =(G1+G2+G3+......+G N ) / N;

[0034] G1 represents the skin glossiness of the first pixel, G2 represents the skin glossiness of the second pixel, and G3 represents the skin glossiness of the third pixel. N The skin glossiness of the Nth pixel;

[0035] The average skin transparency B avg The calculation formula is as follows:

[0036] B avg =(B1+B2+B3+......+B N ) / N;

[0037] B1 represents the skin transparency of the first pixel, B2 represents the skin transparency of the second pixel, and B3 represents the skin transparency of the third pixel. N The skin transparency is the Nth pixel.

[0038] Preferably, based on the custom brightness enhancement value TL, and after simulating brightness enhancement of facial regions, the average skin rosiness R avg Average skin luster (G) avg and average skin transparency B avg This will change, and then the optimized average skin rosiness R will be calculated. avg ’ Optimized average skin radiance G avg ’ And optimized average skin transparency B avg ’ And with average skin rosiness R avg Average skin luster (G) avg and average skin transparency Bavg The calculations were performed to determine the number of noise pixels in the facial region that significantly differed from the average pixel count, as follows:

[0039] According to HZ R =SQRT(R avg -R avg ’ ) 2 HZ G =SQRT(G avg -G avg ’ ) 2 and HZ B =SQRT(B avg -B avg ’ ) 2 The calculation formula for ) is used to calculate the threshold HZ for rosiness difference. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B And calculate the skin rosiness R of the i-th pixel before and after optimization within the total number of pixels N. i Skin gloss G of the i-th pixel i and the skin transparency B of the i-th pixel i A comparison was made one by one, and the differences in rosiness exceeded the threshold HZ. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B The pixels with the largest differences are counted as the number of noise pixels, i.e., the total number of noise pixels XS. noise .

[0040] Preferably, the calculation formula for the optimized facial zone skin smoothness unit is as follows:

[0041] PF = LF × (1 - XS) noise / N)+SQRT(LF×R std ) / 10-XS pore / (LF×2);

[0042] ;

[0043] in:

[0044] PF represents the optimized skin smoothness;

[0045] XS noise Number of noise pixels;

[0046] R std The standard deviation of skin rosiness;

[0047] XSpore This represents the number of pixels in the pores.

[0048] Preferably, the calculation formula for the facial skin uniformity enhancement unit is as follows:

[0049] JF=PF×[1-|R avg -G avg | / (R avg +G avg +1)]+SQRT(LF / (XS pore +1)-PF×(SB / XS pore );

[0050] in:

[0051] JF represents the skin tone uniformity value;

[0052] SB represents the area of ​​the discoloration.

[0053] Preferably, the adjustment steps of the smart beauty method based on the skin tone uniformity value JF are as follows:

[0054] S1. Determine the custom brightness boost value TL according to user needs;

[0055] S2. Calculate the adjusted skin brightness value LF, the optimized skin smoothness value PF, and the skin tone uniformity value JF in sequence.

[0056] S3. Compare the skin tone uniformity value JF with the skin tone uniformity threshold JF. HZ Comparison

[0057] If the skin tone uniformity value JF is higher than the skin tone uniformity threshold JF HZ This indicates that the skin tone in different facial areas is too uniform, easily losing its natural look. Optimization should focus on reducing the area of ​​pores and blemishes to increase the blemish area (SB) and pore pixel count (XS). pore ;

[0058] If the skin tone uniformity value JF is lower than the skin tone uniformity threshold JF HZ This reflects drastic changes in the texture of facial regions. Optimization is applied to increase the area of ​​pores and blemishes to reduce the blemish area (SB) and pore pixel count (XS). pore ;

[0059] If the skin tone uniformity value JF is equal to the skin tone uniformity threshold JF HZ This reflects the effectiveness of the facial contouring treatment when the current facial contouring method is applied.

[0060] Preferably, the database stores skin tone uniformity values ​​JF and skin tone uniformity thresholds JF for different age groups. HZ and regularly adjust the skin tone uniformity threshold JF HZUpdate the skin tone uniformity threshold JF. HZ Specifically, it refers to the average skin tone uniformity value JF for different age groups, and when adjusting and comparing beauty methods, the skin tone uniformity threshold JF for the corresponding age group is extracted. HZ .

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0062] I. This invention improves the accuracy and effect of processing by accurately segmenting the facial region of an image and performing personalized beauty treatments on each sub-region. This is due to the detailed division and targeted processing of different facial regions in the image processing module, which includes a unit for enhancing the brightness of facial skin regions, a unit for optimizing the smoothness of facial skin regions, and a unit for enhancing the uniformity of facial skin regions.

[0063] Second, this invention achieves automated parameter adjustment and scientific calculation by constructing three interrelated algorithmic formulas. Furthermore, in each iteration, based on the skin tone uniformity value JF, it automatically adjusts the pigmentation area SB and the number of pore pixels XS. pore This enables rapid convergence of parameters and the acquisition of the optimal solution.

[0064] Third, this invention can quickly converge to the optimal solution by continuously iterating and adjusting parameters, thereby improving the efficiency of iterative processing. At the same time, by introducing a cyclic influence mechanism, it realizes the cyclic iteration and optimization of the three formulas: the adjusted skin brightness value LF, the optimized skin smoothness value PF, and the skin color uniformity value JF, and further improves the processing effect.

[0065] IV. This invention significantly improves skin tone uniformity by accurately identifying and addressing facial pigmentation and pore issues. In the unit for improving facial skin uniformity in different areas, it introduces pigmentation area SB and pore pixel count XS. pore The parameters enable effective handling of uneven skin tone. Attached Figure Description

[0066] Figure 1 This is a flowchart of the intelligent beauty method based on facial partitioning;

[0067] Figure 2 This is a schematic diagram of the image processing module of the present invention;

[0068] Figure 3 XS is the number of noise pixels in this invention. noise Summary diagram;

[0069] Figure 4 This is a schematic diagram illustrating the adjustment feedback of the skin tone uniformity value JF in this invention. Detailed Implementation

[0070] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0071] This intelligent beauty method based on facial partitioning differs from existing methods, which suffer from insufficient processing precision, difficulty in parameter adjustment, low iteration efficiency, and poor skin tone uniformity. This algorithm unit effectively addresses these shortcomings by employing precise facial partitioning, a scientific parameter adjustment mechanism, an efficient iterative processing flow, and significant improvements in skin tone uniformity. This method not only enhances the processing precision and efficiency of intelligent beauty methods but also provides users with more personalized and precise beauty solutions.

[0072] Example 1, please refer to Figures 1 to 4 This implementation provides a smart beauty method based on facial partitioning, and the specific implementation steps are as follows:

[0073] Step 1: Use the image acquisition module to acquire images of the user's facial regions and transmit them to the image processing module and store them in the database;

[0074] Step 2: Using the image processing module, calculate and output the adjusted skin brightness value LF;

[0075] Step 3: Based on the adjusted skin brightness value LF and the custom brightness enhancement value TL, and using the skin simulation module, simulate and extrapolate the optimized skin state, and summarize the noise pixel count XS. noise ;

[0076] Step 4: Based on the number of noisy pixels XS noise The adjusted skin brightness value LF is calculated, and the optimized skin smoothness PF and skin color uniformity value JF are calculated and output using the image processing module.

[0077] Step 5: Extract the skin tone uniformity threshold JF corresponding to the patient's age group from the database. HZ The skin tone uniformity value (JF) is compared with the skin tone uniformity value (JF), and the beauty method is output by the interactive module.

[0078] The image processing module includes a unit for enhancing the brightness of facial skin regions, a unit for optimizing the smoothness of facial skin regions, and a unit for enhancing the uniformity of facial skin regions.

[0079] The image acquisition module uses equipment including high-resolution cameras and professional photography equipment;

[0080] The equipment used in the image processing module includes a computer and image processing software;

[0081] The equipment used in the skin simulation module includes a skin simulation system;

[0082] The interactive module uses devices including display devices.

[0083] In this embodiment, the system utilizes the cooperation of three algorithm units and combines the results of three calculations—LF, PF, and JF—to form the core of intelligent beauty effect optimization. Specifically, LF is the adjusted skin brightness value, which aims to improve the overall brightness of the skin while considering the influence of shadow areas to achieve a more uniform and bright skin tone effect. PF is the optimized skin smoothness, which aims to reduce blemishes and texture variations on the skin, making the skin look smoother and more delicate. JF is the skin tone uniformity value, which aims to reduce the impact of pigmentation and pores on skin tone, making the skin tone more uniform and natural. Furthermore, the calculation result of JF can also influence the calculation of LF and PF, enabling the three algorithms of this system to achieve comprehensive optimization of skin brightness, smoothness, and skin tone uniformity through continuous iteration and parameter adjustment, thereby providing users with a more personalized and efficient beauty experience.

[0084] Please see Figures 1 to 4 The calculation formula for enhancing facial skin brightness in specific areas is as follows:

[0085] LF=SQRT(R avg +G avg +B avg )×(1+TL / 100)-Y / 2;

[0086] Y = YM / FM;

[0087] in:

[0088] LF represents the adjusted skin brightness value;

[0089] SQRT stands for the radical sign;

[0090] R avg The average skin tone;

[0091] G avg Average skin radiance;

[0092] B avg Average skin transparency;

[0093] TL is a custom brightness boost value. TL reflects the degree to which the user-defined brightness needs to be boosted, and its value ranges from {0-100}.

[0094] Y represents the skin shading intensity;

[0095] YM represents the shadow area, and FM represents the total area of ​​the facial region.

[0096] Average skin rosiness R avg The calculation formula is as follows:

[0097] R avg =(R1+R2+R3+......+R N ) / N;

[0098] N is the total number of pixels, which reflects the total number of pixels used for skin detection within the facial area;

[0099] R1 represents the skin tone of the first pixel, R2 represents the skin tone of the second pixel, R3 represents the skin tone of the third pixel, and R... N The skin rosiness of the Nth pixel;

[0100] Average skin radiance G avg The calculation formula is as follows:

[0101] G avg =(G1+G2+G3+......+G N ) / N;

[0102] G1 represents the skin glossiness of the first pixel, G2 represents the skin glossiness of the second pixel, and G3 represents the skin glossiness of the third pixel. N The skin glossiness of the Nth pixel;

[0103] Average skin transparency B avg The calculation formula is as follows:

[0104] B avg =(B1+B2+B3+......+B N ) / N;

[0105] B1 represents the skin transparency of the first pixel, B2 represents the skin transparency of the second pixel, and B3 represents the skin transparency of the third pixel. N The skin transparency is the Nth pixel.

[0106] In this embodiment: First, in this algorithm unit, "SQRT(R)..." avg +G avg +B avgThe calculation part calculates the square root of the sum of the average values ​​of the red, green, and blue channels of the facial region of the image. This square root of the average value reflects the overall brightness level of the facial region and serves as the basis for the calculation formula of the facial skin brightness unit. It determines the baseline brightness before brightness adjustment. By combining it with the user-defined custom brightness enhancement value TL, the adjusted skin brightness value LF can be calculated.

[0107] The “(1+TL / 100)” calculation part is used to adjust the base brightness according to the user-defined custom brightness increase value TL. By multiplying by the “(1+TL / 100)” calculation part, a linear increase in brightness can be achieved. This calculation part determines the magnitude of the brightness adjustment, that is, how many percentage points the user wants to increase the skin brightness.

[0108] The “Y / 2” calculation part takes into account the influence of the skin shadow degree Y of the facial area of ​​the image on the brightness adjustment. By subtracting the “Y / 2” calculation part, the influence of the shadow area on the overall brightness can be simulated, making the adjusted brightness more natural. This calculation part is used to adjust the brightness to take into account the existence of facial shadows and avoid over-brightening the shadow area.

[0109] This algorithm unit allows users to fine-tune skin brightness according to their aesthetic preferences by allowing them to customize the brightness enhancement value TL. This parameter setting makes beauty processing more flexible and can meet the different brightness needs of different users. Specifically, for users who want brighter skin, the custom brightness enhancement value TL can be increased appropriately, while for users who prefer a natural skin tone, the custom brightness enhancement value TL can be kept at a lower level.

[0110] In addition, the introduction of skin shadow degree Y enables the facial partition skin brightness unit to more accurately locate and process shadow areas in the face. By increasing the brightness of these areas, the overall brightness uniformity of the face can be significantly improved, and the shadow effect caused by uneven light can be reduced. This treatment not only improves the visual effect of the skin, but also makes the facial contour clearer and enhances the three-dimensionality.

[0111] The calculation of the average values ​​of the red, green, and blue channels ensures color balance. During processing, enhancing the skin brightness unit in the facial area ensures that the brightness changes between these three channels remain coordinated, avoiding color distortion caused by adjusting a single channel. This color balance processing makes the skin look more natural and healthy.

[0112] Please see Figures 1 to 4 The calculation formula for optimizing the facial skin smoothness unit is as follows:

[0113] PF = LF × (1 - XS) noise / N)+SQRT(LF×R std ) / 10-XS pore / (LF×2);

[0114] ;

[0115] in:

[0116] PF represents the optimized skin smoothness;

[0117] XS noise Number of noise pixels;

[0118] R std The standard deviation of skin rosiness;

[0119] XS pore This represents the number of pixels in the pores.

[0120] In this embodiment, firstly, "LF×(1-XS)" noise The / N) calculation section calculates the number of noisy pixels XS. noise The ratio relative to the total number of pixels N is used to adjust the adjusted skin brightness value LF. By reducing the influence of noise pixels on the adjusted skin brightness value LF, the smoothness of the skin can be optimized. It determines the degree of influence of noise on smoothness optimization, and by reducing noise interference, the skin looks smoother.

[0121] “SQRT(LF×R std The calculation of ) / 10” takes into account the standard deviation R of skin redness in the red channel of the facial region of the image. std The degree of texture change is calculated by multiplying the adjusted skin brightness value LF and dividing by 10. Here, 10 is a weighting coefficient used to adjust the degree of influence of texture change on smoothness, thereby simulating the effect of texture change on smoothness. It is used to consider texture change in smoothness optimization to avoid over-smoothing that causes the skin to lose details.

[0122] XS pore Even if the pixel count of pores is partially considered, / (LF×2)” pore The effect on smoothness is measured by the number of pixels per pore (XS). pore Dividing by the adjusted skin brightness value LF multiplied by 2 can reduce the negative impact of pores on smoothness. Here, 2 is a weighting coefficient used to adjust the degree of influence of pores on smoothness. This calculation part is used to reduce the interference of pores in smoothness optimization, making the skin look more delicate.

[0123] This algorithm unit uses the number of noise pixels XS noiseThe ratio of the number of pixels to the total number of pixels N reflects the density of noise in the image. Optimizing the facial region's skin smoothness unit, by calculating this ratio, can accurately identify and suppress noise in the image. This processing not only reduces skin blemishes but also makes the skin look smoother and more delicate. The standard deviation R of skin rosiness is also measured. std Reflecting changes in skin texture, optimizing the facial skin smoothness unit by adjusting this parameter can improve the skin texture effect, making it look more natural and even. This texture optimization process makes the skin appear more delicate and elastic, while reducing the number of pore pixels (XS). pore This design allows the optimized facial skin smoothness unit to refine pores. By reducing the number of pixels in the pores, it can significantly improve the smoothness of the skin, making it look firmer and smoother. This is undoubtedly a great benefit for users who pursue delicate skin.

[0124] Please see Figures 1 to 4 The calculation formula for improving the uniformity of facial skin in different zones is as follows:

[0125] JF=PF×[1-|R avg -G avg | / (R avg +G avg +1)]+SQRT(LF / (XS pore +1)-PF×(SB / XS pore );

[0126] in:

[0127] JF represents the skin tone uniformity value;

[0128] SB represents the area of ​​the discoloration.

[0129] In this embodiment, the algorithm unit first performs "PF×[1-|R avg -G avg | / (R avg +G avg The calculation section considers the ratio of the difference between the average values ​​of the red and green channels to the sum of the two, which is used to simulate the degree of skin unevenness. By adjusting this ratio, the skin uniformity can be optimized. This calculation section determines the impact of the degree of skin unevenness on the improvement of uniformity. By reducing the negative impact of the difference between the red and green channels on the skin uniformity value JF, the skin is made more uniform.

[0130] “SQRT(LF / (XS pore +1) The calculation takes into account the adjusted skin brightness value LF relative to the number of pore pixels XS. poreThe ratio is calculated to further adjust the evenness of skin tone, especially for the pore area. This calculation part is used to take into account the influence of pores in the evenness improvement, and makes the skin look more even by adjusting the brightness difference in the pore area.

[0131] “PF×(SB / XS pore The calculation part considers the impact of the pigmentation area SB on skin tone uniformity by subtracting the optimized skin smoothness PF multiplied by the pigmentation area SB and the number of pore pixels XS. pore The ratio of pigmentation can reduce the negative impact of pigmentation on skin tone evenness. It is used to reduce the interference of pigmentation in the evenness improvement, making the skin tone look more even and consistent.

[0132] This algorithm unit can directly affect the hue and saturation of skin tone by adjusting the average values ​​of the red and green channels, and improve the skin uniformity of facial regions. By calculating these two parameters and adjusting them appropriately, the unit can significantly improve the uniformity and naturalness of skin tone. This skin tone correction process makes the skin look healthier and more radiant.

[0133] The consideration of the pigmentation area SB enables the facial skin uniformity enhancement unit to lighten pigmentation. By calculating the pigmentation area SB and making appropriate adjustments, the impact of pigmentation on skin tone can be significantly reduced, and skin tone uniformity can be improved. This treatment is undoubtedly an effective solution for users with pigmentation problems.

[0134] In addition, the facial skin uniformity enhancement unit not only considers the treatment of skin brightness, smoothness and pigmentation fading, but also achieves comprehensive optimization of skin tone by integrating the results of these parameters. This comprehensive optimization makes the skin look more uniform and natural, thus improving the overall beauty effect.

[0135] In summary, the beneficial effects of the facial zone skin brightness enhancement unit, the facial zone skin smoothness optimization unit, and the facial zone skin uniformity enhancement unit are all closely integrated with their parameters and functions. The setting and adjustment of these parameters make beauty treatments more precise and flexible, and can meet the different skin beauty needs of different users.

[0136] Please see Figures 1 to 4 The adjustment steps for the intelligent beauty method based on skin tone evenness value JF are as follows:

[0137] S1. Determine the custom brightness boost value TL according to user needs;

[0138] S2. Calculate the adjusted skin brightness value LF, the optimized skin smoothness value PF, and the skin tone uniformity value JF in sequence.

[0139] S3. Compare the skin tone uniformity value JF with the skin tone uniformity threshold JF. HZ Comparison

[0140] If the skin tone uniformity value JF is higher than the skin tone uniformity threshold JF HZ This indicates that the skin tone in different facial areas is too uniform, easily losing its natural look. Optimization should focus on reducing the area of ​​pores and blemishes to increase the blemish area (SB) and pore pixel count (XS). pore ;

[0141] If the skin tone uniformity value JF is lower than the skin tone uniformity threshold JF HZ This reflects drastic changes in the texture of facial regions. Optimization is applied to increase the area of ​​pores and blemishes to reduce the blemish area (SB) and pore pixel count (XS). pore ;

[0142] If the skin tone uniformity value JF is equal to the skin tone uniformity threshold JF HZ This reflects the effectiveness of the facial contouring treatment when the current facial contouring method is applied.

[0143] In this embodiment, the algorithm unit can dynamically adjust the pigmentation area SB and the number of pore pixels XS based on the skin tone uniformity value J calculated by the facial skin uniformity enhancement unit. pore This makes skin brightness adjustment more precise. Through multiple iterations, the results of improving the skin brightness unit and the skin evenness unit in the facial area will gradually converge to achieve a satisfactory level of skin tone evenness, avoiding the problems of over-adjustment or under-adjustment. Moreover, the cyclical influence mechanism enables the skin brightness unit and the skin evenness unit in the facial area to cooperate with each other to achieve overall optimization of skin brightness, smoothness and skin tone evenness, thereby improving the effect of smart beauty methods and user experience.

[0144] Furthermore, when the custom brightness enhancement value TL is defined too high, resulting in uneven facial zoning in the skin tone uniformity value JF, the parameters in the formula can be dynamically adjusted by adjusting the shadow area hand grip ratio, optimizing noise processing, and adjusting the number of pore pixels and the area of ​​pigmentation SB strategies, thereby optimizing skin tone uniformity. This adjustment is based on real-time feedback, which enables precise processing of specific problems. Through a cyclic feedback mechanism, the algorithm can continuously receive processed image data and adjust parameters according to the display results of the skin tone uniformity value JF. This continuous optimization process can gradually approach the ideal skin tone uniformity state, improving the accuracy and reliability of the algorithm.

[0145] During the adjustment process, the algorithm considers multiple factors to avoid over-processing that results in an overly uniform skin tone and a loss of naturalness. By finely adjusting parameters, the algorithm can improve skin tone uniformity while maintaining the skin's natural texture and details. Specifically, if the skin tone uniformity value (JF) is too high, the algorithm will attempt to increase the area of ​​pigmentation spots (SB) and the number of pore pixels (XS). pore The value is adjusted to restore the natural unevenness of skin tone, which helps maintain the skin's vibrancy and realism.

[0146] The parameter adjustment strategy in the algorithm can adapt to skin conditions of different ages, skin types and under different lighting conditions. By flexibly adjusting the parameters, the algorithm can accurately process various skin conditions, improve the adaptability of the algorithm, and maintain stable performance when facing various complex situations. This robustness helps the algorithm cope with various challenges and uncertainties in practical applications.

[0147] Through a feedback loop, the algorithm can customize personalized beauty solutions based on the user's personal preferences and skin condition. This personalized solution can more accurately meet the user's needs, improve user satisfaction and loyalty, and the feedback loop allows the algorithm to receive user feedback in real time and make adjustments. This interactivity helps to enhance the user's sense of participation and trust, and improve the overall level of user experience.

[0148] In summary, by combining the facial region skin brightness enhancement unit, the facial region skin smoothness unit, and the facial region skin uniformity unit and their parameters, the cyclical feedback mechanism formed by the above adjustment strategies has brought significant benefits in improving the accuracy of skin tone uniformity value JF, maintaining the naturalness of the skin, improving the algorithm's adaptability and robustness, and enhancing the user experience. Moreover, this mechanism enables the algorithm to process skin image data more accurately, providing users with more personalized, efficient, and reliable beauty services.

[0149] Example 2, please refer to Figures 1 to 4 Based on a custom brightness enhancement value TL, and after simulating brightness enhancement on facial areas, the average skin rosiness R... avg Average skin luster (G) avg and average skin transparency B avg This will change, and then the optimized average skin rosiness R will be calculated. avg ’ Optimized average skin radiance G avg ’ And optimized average skin transparency B avg ’ And with average skin rosiness R avg Average skin luster (G) avg and average skin transparency Bavg The calculations were performed to determine the number of noise pixels in the facial region that significantly differed from the average pixel count, as follows:

[0150] According to HZ R =SQRT(R avg -R avg ’ ) 2 HZ G =SQRT(G avg -G avg ’ ) 2 and HZ B =SQRT(B avg -B avg ’ ) 2 The calculation formula for ) is used to calculate the threshold HZ for rosiness difference. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B And calculate the skin rosiness R of the i-th pixel before and after optimization within the total number of pixels N. i Skin gloss G of the i-th pixel i and the skin transparency B of the i-th pixel i A comparison was made one by one, and the differences in rosiness exceeded the threshold HZ. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B The pixels with the largest differences are counted as the number of noise pixels, i.e., the total number of noise pixels XS. noise .

[0151] In this embodiment, the average skin redness R in the original image data avg Average skin luster (G) avg and average skin transparency B avg These are important parameters describing the color characteristics of facial regions in an image. These average values ​​reflect the overall brightness and intensity of facial skin in different color channels. When the skin simulation module simulates and calculates the differences in the optimized skin state, it specifically compares image data before and after optimization. Specifically, it calculates the average redness R of the optimized facial region. avg ’ Optimized average skin radiance G avg ’ And optimized average skin transparency B avg ’ The difference is then compared with the original value, and the resulting difference can quantitatively describe the impact of the optimization algorithm on facial skin color characteristics. This simulation method can also help advance and present users' beauty methods.

[0152] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart beauty method based on facial partitioning, characterized in that, The specific implementation steps are as follows: Step 1: Use the image acquisition module to acquire images of the user's facial regions and transmit them to the image processing module and store them in the database; Step 2: Using the image processing module, calculate and output the adjusted skin brightness value LF; Step 3: Based on the adjusted skin brightness value LF and the custom brightness enhancement value TL, and using the skin simulation module, simulate and extrapolate the optimized skin state, and summarize the noise pixel count XS. noise ; Step 4: Based on the number of noisy pixels XS noise The adjusted skin brightness value LF is calculated, and the optimized skin smoothness PF and skin color uniformity value JF are calculated and output using the image processing module, respectively. Step 5: Extract the skin tone uniformity threshold JF corresponding to the patient's age group from the database. HZ The skin tone uniformity value (JF) is compared with the skin tone uniformity value (JF), and the beauty method is output by the interactive module. The image processing module includes a unit for enhancing the brightness of facial skin regions, a unit for optimizing the smoothness of facial skin regions, and a unit for enhancing the uniformity of facial skin regions. The calculation formula for enhancing facial skin brightness in specific areas is as follows: LF=SQRT(R avg +G avg +B avg )×(1+TL / 100)-Y / 2; Y = YM / FM; in: LF represents the adjusted skin brightness value; SQRT stands for the radical sign; R avg The average skin tone; G avg Average skin radiance; B avg Average skin transparency; TL is a custom brightness boost value. TL reflects the degree to which the user-defined brightness needs to be boosted, and its value ranges from {0-100}. Y represents the skin shading intensity; YM represents the shadow area, and FM represents the total area of ​​the facial region. The calculation formula for the optimized facial zone skin smoothness unit is as follows: PF=LF×(1-XS noise / N)+SQRT(LF×R std ) / 10-XS pore / (LF×2); ； in: PF represents the optimized skin smoothness; XS noise Number of noise pixels; R std The standard deviation of skin rosiness; XS pore Number of pixels for pores; The calculation formula for the facial skin uniformity enhancement unit is as follows: JF=PF×[1-|R avg -G avg | / (R avg +G avg +1)]+SQRT(LF / (XS pore +1)-PF×(SB / XS pore ); in: JF represents the skin tone uniformity value; SB represents the area of ​​the discoloration.

2. The intelligent beauty method based on facial partitioning according to claim 1, characterized in that, The image acquisition module includes a high-resolution camera and professional photography equipment; The image processing module includes a computer and image processing software; The skin simulation module includes a skin simulation system; The interactive module includes a display device.

3. The intelligent beauty method based on facial partitioning according to claim 2, characterized in that: The average skin rosiness R avg The calculation formula is as follows: R avg =(R1+R2+R3+......+R N ) / N; N is the total number of pixels, which reflects the total number of pixels used for skin detection within the facial area; R1 represents the skin tone of the first pixel, R2 represents the skin tone of the second pixel, R3 represents the skin tone of the third pixel, and R... N The skin rosiness of the Nth pixel; The average skin luster G avg The calculation formula is as follows: G avg =(G1+G2+G3+......+G N ) / N; G1 represents the skin glossiness of the first pixel, G2 represents the skin glossiness of the second pixel, and G3 represents the skin glossiness of the third pixel. N The skin glossiness of the Nth pixel; The average skin transparency B avg The calculation formula is as follows: B avg =(B1+B2+B3+......+B N ) / N; B1 represents the skin transparency of the first pixel, B2 represents the skin transparency of the second pixel, and B3 represents the skin transparency of the third pixel. N The skin transparency is the Nth pixel.

4. The intelligent beauty method based on facial partitioning according to claim 3, characterized in that: Based on the custom brightness enhancement value TL, and after simulating brightness enhancement on facial areas, the average skin rosiness R avg Average skin luster (G) avg and average skin transparency B avg This will change, and then the optimized average skin rosiness R will be calculated. avg ’ Optimized average skin radiance G avg ’ And optimized average skin transparency B avg ’ And with average skin rosiness R avg Average skin luster (G) avg and average skin transparency B avg The calculations were performed to determine the number of noise pixels in the facial region that significantly differed from the average pixel count, as follows: According to HZ R =SQRT(R avg -R avg ’ ) 2 HZ G =SQRT(G avg -G avg ’ ) 2 and HZ B =SQRT(B avg -B avg ’ ) 2 The calculation formula for ) is used to calculate the threshold HZ for rosiness difference. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B And calculate the skin rosiness R of the i-th pixel before and after optimization within the total number of pixels N. i Skin gloss G of the i-th pixel i and the skin transparency B of the i-th pixel i A comparison was made one by one, and the differences in rosiness exceeded the threshold HZ. R Gloss difference threshold (HZ) G and transparency difference threshold HZ B The pixels with the largest differences are counted as the number of noise pixels, i.e., the total number of noise pixels XS. noise .

5. The intelligent beauty method based on facial partitioning according to claim 4, characterized in that: The adjustment steps for the intelligent beauty method based on the skin tone uniformity value JF are as follows: S1. Determine the custom brightness boost value TL according to user needs; S2. Calculate the adjusted skin brightness value LF, the optimized skin smoothness value PF, and the skin tone uniformity value JF in sequence. S3. Compare the skin tone uniformity value JF with the skin tone uniformity threshold JF. HZ Comparison If the skin tone uniformity value JF is higher than the skin tone uniformity threshold JF HZ This indicates that the skin tone in different facial areas is too uniform, easily losing its natural look. Optimization should focus on reducing the area of ​​pores and blemishes to increase the blemish area (SB) and pore pixel count (XS). pore ; If the skin tone uniformity value JF is lower than the skin tone uniformity threshold JF HZ This reflects drastic changes in the texture of facial regions. Optimization is applied to increase the area of ​​pores and blemishes to reduce the blemish area (SB) and pore pixel count (XS). pore ; If the skin tone uniformity value JF is equal to the skin tone uniformity threshold JF HZ This reflects the effectiveness of the facial contouring treatment when the current facial contouring method is applied.

6. The intelligent beauty method based on facial partitioning according to claim 5, characterized in that: The database stores skin tone uniformity values ​​JF and skin tone uniformity thresholds JF for different age groups. HZ and regularly adjust the skin tone uniformity threshold JF HZ Update the skin tone uniformity threshold JF. HZ Specifically, it refers to the average skin tone uniformity value JF for different age groups, and when adjusting and comparing beauty methods, the skin tone uniformity threshold JF for the corresponding age group is extracted. HZ .

Images