Full-automatic hair dyeing and color matching machine and user personalized recommendation system and method

Abstract

The invention relates to the technical field of personalized recommendation, and particularly discloses a full-automatic hair dyeing and color matching machine and a user personalized recommendation system and method, and the method comprises the following steps: obtaining a user image; extracting user skin color features based on the user image; performing illumination correction processing on the skin color features of the user to obtain reference skin color features under standard illumination; selecting a plurality of candidate hair colors from a hair color database based on the reference skin color features; acquiring a plurality of pieces of predefined illumination scene data; simulating and generating a visual representation color value of each candidate hair color based on predefined illumination scene data; evaluating a comprehensive adaptability index of each candidate hair color based on the visual representation color value; selecting at least one recommended hair color from the candidate hair colors according to the comprehensive adaptability index; and outputting the recommended hair color. According to the method, the adaptability and the stability of the hair dyeing effect in different life scenes are improved, and the user experience satisfaction degree is enhanced.

Description

Technical Field

[0001] This invention relates to the field of personalized recommendation technology, specifically to a fully automatic hair dyeing and color matching machine and a user personalized recommendation system and method. Background Technology

[0002] A color mixing machine is a professional device used for automatically mixing hair dye. Its hardware system is highly integrated and intelligently operated, enabling efficient and precise color mixing. It mainly consists of three parts: an LED touchscreen display, serving as the core interface for human-machine interaction, allowing users to intuitively view the operating status, select functions, and make settings, achieving real-time monitoring and control of the entire color mixing process; a dye bottle rotation and selection system, which, driven by a motor, automatically and precisely rotates the dye bottle of the specified color to the dispensing position according to instructions, enabling rapid switching and recall of multiple colors, greatly improving color mixing efficiency; and a precision metering and pumping system, combined with high-precision weighing sensors and a pump assembly structure, responsible for the precise metering and delivery of the dye and developer, ensuring that the error in the amount of each colorant dispensed is controlled within a minimal range, thereby guaranteeing that each color mixing achieves the desired color effect.

[0003] When using color matching tools to recommend personalized hair colors, it's typically necessary to analyze the user's skin tone in their photos to suggest suitable hair colors, and then use AI technology to visually replace the hair color in the image for a preview. However, this method has a problem: photos are often taken under specific lighting conditions, while users encounter various lighting scenarios in their daily lives. This can cause the recommended hair color to look ideal in a photo but perform poorly in actual natural light, indoor lighting, or other environments, resulting in unsatisfactory hair coloring results in certain situations. Summary of the Invention

[0004] The purpose of this invention is to provide a fully automatic hair dyeing and color matching machine and a personalized recommendation system and method for users, thereby solving the above-mentioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A personalized recommendation method for a fully automatic hair dyeing and color matching machine includes the following steps:

[0007] Acquire user image; extract user skin color features based on user image;

[0008] The user's skin color features are subjected to illumination correction processing to obtain reference skin color features under standard illumination;

[0009] Multiple candidate hair colors are selected from the hair color database based on the reference skin color features;

[0010] Acquire multiple predefined lighting scene data, each representing a typical ambient lighting condition; simulate and generate the visual representation color value of each candidate color based on the predefined lighting scene data.

[0011] The overall adaptability index of each candidate hair color under all predefined lighting scene data is evaluated based on the visual performance color value; at least one recommended hair color is selected from the candidate hair colors according to the overall adaptability index; and the recommended hair color is output.

[0012] Preferably, extracting user skin color features includes:

[0013] Facial regions are identified from user images by analyzing contours and color distribution in the image; skin color pixel sets are located within facial regions, and if a pixel's color value belongs to a preset skin color range, then the pixel belongs to the skin color pixel set.

[0014] The average color value is calculated as the user's skin color feature from the set of skin color pixels. The average color value is obtained by summing the color components of all skin color pixels and dividing by the number of pixels. The color components include red, green and blue components.

[0015] Preferably, the illumination correction process includes:

[0016] The illumination condition parameters of the user image are obtained by analyzing the highlight and shadow regions in the image. The highlight and shadow regions are determined based on the changes in pixel brightness values.

[0017] The user's skin color features are adjusted to the color value under standard lighting conditions based on the lighting condition parameters. The standard lighting environment is defined as uniform white light conditions. The adjustment process is achieved through a color conversion function, which is built based on the lighting physical model. The lighting physical model takes into account the influence of lighting color temperature and intensity.

[0018] The adjusted user skin tone features are used as reference skin tone features.

[0019] Preferably, selecting multiple candidate hair colors includes:

[0020] Calculate the color similarity between the reference skin color feature and each hair color option in the hair color database. The color similarity is evaluated in a color space using Euclidean distance, and the color space is represented using RGB or LAB.

[0021] The hair color options are sorted in descending order of color similarity, and the top-ranked hair colors are selected as candidate hair colors. The number of selections is based on a pre-set manual setting.

[0022] Preferably, the visual representation color value of each candidate hair color is generated by simulating predefined lighting scene data, including:

[0023] Based on the lighting color and intensity parameters contained in the predefined lighting scene data, the original color values ​​of the candidate hair colors are subjected to lighting transformation. The lighting transformation is achieved through a color mixing process, which simulates the physical interaction between lighting and hair color, and superimposes the effect of the lighting color and intensity parameters onto the original color values ​​of the candidate hair colors. The color value output after the color mixing process is the visual representation color value of the candidate hair colors under the predefined lighting scene data.

[0024] Preferably, the comprehensive adaptability metrics for evaluating each candidate hair color include:

[0025] For each candidate hair color, calculate the color difference between its visual representation color value under all predefined lighting scene data and the color difference value of the reference skin color feature. The color difference value is calculated in the color space using the color distance formula.

[0026] The color difference values ​​of all predefined lighting scene data are weighted and averaged. The weighted average result is used as a comprehensive adaptability index. The lower the comprehensive adaptability index value, the more stable the coordination between hair color and skin color under different lighting scenes. The weights are set based on the usage frequency data of the environment represented by each lighting scene data. The usage frequency data is obtained from user environment surveys.

[0027] Preferably, selecting at least one recommended hair color from the candidate hair colors includes:

[0028] Candidate hair colors are sorted according to the comprehensive adaptability index, from the lowest to the highest comprehensive adaptability index; the top-ranked candidate hair colors are selected as recommended hair colors, and the number of selections is a pre-set number.

[0029] A personalized recommendation system for a fully automatic hair dyeing and color matching machine includes:

[0030] Feature acquisition module: Acquires user image; extracts user skin color features based on the user image;

[0031] The user's skin color features are subjected to illumination correction processing to obtain reference skin color features under standard illumination;

[0032] Data processing module: Selects multiple candidate hair colors from the hair color database based on the reference skin color features;

[0033] Acquire multiple predefined lighting scene data, each representing a typical ambient lighting condition; simulate and generate the visual representation color value of each candidate color based on the predefined lighting scene data.

[0034] Recommendation module: Evaluates the comprehensive adaptability index of each candidate hair color under all predefined lighting scene data based on the visual performance color value; selects at least one recommended hair color from the candidate hair colors according to the comprehensive adaptability index; and outputs the recommended hair color.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] 1) This invention, by introducing a lighting correction process, effectively eliminates the interference of specific shooting environment lighting on skin color features in the user's original image, thereby extracting more realistic and accurate reference skin color features under standard lighting. This step ensures that the base skin color for subsequent hair color recommendations is not affected by accidental shooting conditions, making the recommended hair color based on the essential attributes of the user's skin color rather than temporary lighting effects. This significantly improves the accuracy and inherent harmony between hair color and the user's own skin color, making the recommendation results more reliable and personalized.

[0037] 2) This invention simulates and comprehensively evaluates the visual performance of candidate hair colors under multiple predefined typical lighting scenarios, effectively predicting and ensuring the comprehensive performance of recommended hair colors in different daily lighting environments (such as natural light, indoor lighting, etc.). This method overcomes the limitations of recommendation results under a single lighting condition, ensuring that the selected recommended hair color maintains good coordination and visual appeal with the user's skin tone in various common lighting environments, thereby significantly improving the adaptability and stability of hair dyeing effects in different life scenarios and enhancing user experience satisfaction. Attached Figure Description

[0038] The invention will now be further described with reference to the accompanying drawings.

[0039] Figure 1 This is a schematic diagram of the overall operation of a hair dyeing and color mixing machine;

[0040] Figure 2 This is a schematic diagram of the internal structure of a hair dyeing and color mixing machine;

[0041] Figure 3 This is a flowchart illustrating a user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to the present invention.

[0042] In the picture: 1. Color mixing machine casing; 2. Receiving bowl; 3. LED touch screen display; 4. Dye bottle. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please see Figures 1-2 As shown, this invention is applicable to a hair dyeing and color mixing machine. The hair dyeing and color mixing machine mainly includes a transparent outer shell 1, on which an LED touch screen 3 is installed. The LED touch screen 3 is used to display recommended hair colors for user selection. Inside the transparent outer shell, there are several dye bottles 4. Different dye bottles 4 are used to store different colors of hair dye. A receiving bowl 2 is installed below the dye bottles 4 to receive the hair dye pumped from the dye bottles 4. After the user selects their desired hair color, the hair dyeing process can begin. According to the user's selected hair color, different dye bottles 4 start dispensing dye to the receiving bowl. During the dispensing process, the dye bottle rotation and selection system drives a turntable via a motor to automatically and precisely rotate the dye bottle 4 of the specified color to the receiving bowl 2. After one dye bottle 4 is dispensed, the system controls a new dye bottle 4 to rotate to the receiving bowl 2. At the same time, the amount dispensed is controlled by a precision metering and pumping system to ensure that the dosage error of each type of hair dye is controlled within a very small range, ensuring that each color adjustment achieves the expected color effect. After dispensing, the hair dye in the receiving bowl 2 is mixed to obtain the hair dye required for the user's selected hair color.

[0045] Please see Figure 3 As shown, this invention provides a personalized recommendation method for a fully automatic hair dyeing and color matching machine, comprising the following steps:

[0046] Acquire user images; extract user skin color features based on the user images.

[0047] In a preferred embodiment of the present invention, extracting user skin color features includes:

[0048] Facial regions are identified from user images by analyzing contours and color distribution in the image. Within each facial region, a set of skin-colored pixels is located. If a pixel's color value falls within a preset skin-colored range, then the pixel belongs to the skin-colored pixel set.

[0049] The average color value is calculated as the user's skin color feature from the set of skin color pixels. The average color value is obtained by summing the color components of all skin color pixels and dividing by the number of pixels. The color components include red, green and blue components.

[0050] For example, the process of identifying facial regions from a user image is based on a comprehensive analysis of the structural and color distribution features of the facial regions in the image. After the image is input, it first undergoes preliminary recognition by a face detection module. Face detection relies on feature extraction algorithms, which analyze information such as pixel brightness, color gradient, and edge direction in the image to determine whether there are regions that match the distribution of facial features. After detecting potential candidate face regions, the algorithm further uses facial feature point localization methods to calibrate key facial positions, including the positions of typical features such as the eyes, bridge of the nose, lips, and chin. The positional relationships of these feature points form the basis of the facial contour. Based on the spatial distribution of these feature points, the circumscribed boundary region of the face is calculated, and this region is used as the initial range of the facial regions. To improve the accuracy of recognition, the boundaries of the initial facial regions are refined by analyzing the continuity of pixel color changes inside and outside the region to determine the true boundaries of the skin region, avoiding misidentification of hair, clothing, or background as part of the face.

[0051] After determining the facial regions, statistical analysis is performed on the color distribution within each region to identify the skin tone pixel set. Skin tone recognition is based on a color feature thresholding method. In the RGB color space, skin tone typically has a certain hue range and brightness characteristics. The color value of each pixel is evaluated; if the proportions of its red, green, and blue components conform to a preset skin tone color range standard, and its brightness value is within a reasonable range, then the pixel is considered to belong to the skin tone pixel set. To avoid misjudgments caused by lighting intensity or camera color cast, the overall image brightness is normalized, making the skin tone recognition process more stable. After the skin tone pixel set is established, the color components of all pixels within the set are accumulated, and the average color value is calculated based on the total number of pixels. The average color value is not dependent on any specific model but is obtained by statistically averaging the actual color distribution of skin tone pixels, thus forming a numerical expression that represents the user's overall skin tone characteristics. This average color value is highly stable and less affected by individual abnormal pixels, accurately reflecting the user's true skin tone.

[0052] The user's skin color features are subjected to illumination correction processing to obtain reference skin color features under standard illumination.

[0053] In another embodiment of the present invention, the illumination correction process includes:

[0054] The illumination condition parameters of the user image are obtained by analyzing the highlight and shadow regions in the image. The highlight and shadow regions are determined based on the changes in pixel brightness values.

[0055] The user's skin color features are adjusted to the color value under standard lighting conditions based on the lighting condition parameters. The standard lighting environment is defined as uniform white light conditions. The adjustment process is achieved through a color conversion function, which is built based on a lighting physics model that takes into account the effects of light color temperature and intensity.

[0056] The adjusted user skin tone features are used as reference skin tone features.

[0057] Specifically, the process of performing illumination correction on user skin tone features involves analyzing the image's illumination conditions and restoring the original skin tone features to their color representation under standard lighting conditions to ensure the authenticity and reliability of the skin tone features. Obtaining illumination condition parameters relies on analyzing the brightness distribution of the user image. Different regions in the image exhibit significant brightness differences under illumination. By statistically analyzing and judging these brightness changes, key indicators reflecting illumination characteristics can be extracted: After input, the image first undergoes brightness channel separation processing to extract the brightness value information of each pixel. Then, through brightness histogram analysis, the distribution range of bright and dark areas is identified. Bright areas typically correspond to the facial areas directly illuminated by the light source, while dark areas correspond to shadow areas. The difference in brightness between the two reflects the distribution characteristics of light intensity. Illumination condition parameters include three main components: light intensity, light direction, and color temperature. Light intensity is estimated through the brightness contrast between bright and shadow areas; light direction is inferred from the brightness gradient change trend of different areas; and color temperature is determined based on the average color bias of the bright areas. When the light color leans towards warm tones, it indicates a lower color temperature; conversely, a cooler color tone indicates a higher color temperature. These parameters allow for a relatively accurate description of the color attributes and illumination patterns of the light during photography.

[0058] After obtaining the lighting condition parameters, they need to be applied to the skin tone feature adjustment process. The adjustment process uses a standard lighting environment as a reference, defined as uniformly distributed white light conditions, meaning the light source has no directional deviation in space and a neutral color temperature. During correction, a color conversion function maps the user's skin tone features, converting their color values ​​from the original lighting environment to their representation under standard lighting. The color conversion function is built based on a lighting physics model, which describes the reflection and absorption characteristics of light after it strikes an object's surface. The model considers the interaction between the color temperature of the light source, the intensity of the illumination, and the reflectivity of the skin surface. Using this model, the changing trends of color channels under different lighting conditions can be calculated, thus deriving the color conversion relationship from the original lighting environment to the standard lighting environment.

[0059] During the conversion process, each color component of the skin tone feature is independently adjusted to achieve a balanced state under standard conditions across all three channels, eliminating deviations caused by changes in light intensity or color temperature. The resulting color value is the reference skin tone feature under standard lighting. This reference skin tone feature accurately reflects the user's skin tone information under natural neutral lighting, avoiding distortion caused by lighting deviations. The principle behind the entire lighting correction process lies in inferring the lighting characteristics of the shooting environment through physical analysis of image brightness and color distribution, and then reversing the color based on a physical lighting model. This makes the expression of skin tone features independent of the external lighting environment, providing a unified and reliable skin tone data foundation for subsequent hair color recommendations.

[0060] Multiple candidate hair colors are selected from a hair color database based on reference skin color features.

[0061] In another embodiment of the present invention, selecting multiple candidate hair colors includes:

[0062] The color similarity between the reference skin color feature and each hair color option in the hair color database is calculated. The color similarity is evaluated in a color space using Euclidean distance, and the color space is represented using RGB or LAB.

[0063] The hair color options are sorted in descending order of color similarity, and the top-ranked hair colors are selected as candidate hair colors. The number of selections is based on a pre-set manual setting.

[0064] For example, in the process of selecting multiple candidate hair colors, the reference skin tone features are first compared with the color information of each hair color option in the hair color database. To achieve comparison between color features, the skin tone features and hair color data need to be mapped to the same color space. This color space can be the RGB space with red, green, and blue components, or the LAB space with lightness, chroma, and color difference as features. By measuring the differences in the component values ​​of each color sample in this color space, the color similarity between the reference skin tone and each hair color sample is obtained. The color similarity is evaluated based on the distance relationship between each color component, which reflects the visual proximity of the two colors; the smaller the distance, the closer the colors are. After calculating the similarity for all hair color options, the similarity results of all hair color samples are sorted based on the magnitude of similarity, with hair colors closer to the reference skin tone ranking higher. Based on this ranking, several top-ranked hair colors are selected as candidate hair colors according to a preset threshold. The number of candidate hair colors can be set according to system settings or human experience to ensure a balance between the computational load and recommendation accuracy in subsequent lighting adaptability assessments.

[0065] Acquire multiple predefined lighting scene data, each representing a typical ambient lighting condition; simulate and generate the visual representation color value of each candidate color based on the predefined lighting scene data.

[0066] In another embodiment of the present invention, the visual representation color value of each candidate color is generated by simulating predefined lighting scene data, including:

[0067] Based on the lighting color and intensity parameters contained in the predefined lighting scene data, the original color values ​​of the candidate hair colors are subjected to lighting transformation. The lighting transformation is achieved through a color mixing process, which simulates the physical interaction between lighting and hair color, and superimposes the effect of the lighting color and intensity parameters onto the original color values ​​of the candidate hair colors. The color value output after the color mixing process is the visual representation color value of the candidate hair colors under the predefined lighting scene data.

[0068] Understandably, predefined lighting scene data is a set of parameters used to simulate typical lighting conditions in real-world environments, reproducing the visual representation of colors under different lighting conditions in recommendation calculations. Each set of lighting scene data includes elements such as light color, light intensity, light direction, and environmental reflectivity, used to characterize the lighting features in a specific scene. The settings of the predefined lighting scene data are determined based on common user environments, which are established as standard models through long-term measurements and spectral analysis of actual lighting conditions. Typical scenarios include outdoor environments under natural sunlight, indoor incandescent lighting in homes, cool-colored fluorescent lighting in offices, soft diffused overcast lighting environments, and mixed lighting commercial space environments. The lighting parameters for each scene are set according to the spectral distribution of the light source, illuminance level, and reflectivity characteristics to ensure that they represent universally applicable lighting conditions.

[0069] When simulating the visual representation of candidate hair colors, predefined lighting scene data is used as input to perform lighting transformation on the original color values ​​of the candidate hair colors. The principle of lighting transformation is to reconstruct the visual representation of hair color under specific lighting conditions through a color mixing model. The original color values ​​of the candidate hair colors represent color data measured under standard lighting. When lighting color and intensity parameters are introduced, the interaction between light and hair color needs to be calculated to superimpose the lighting effect onto the original color values. The color mixing process is not a simple numerical weighting, but a simulation method based on the physical characteristics of lighting. By comprehensively calculating the incident angle of light, the reflection component and transmission component of the hair surface, and the ambient diffuse reflection, the color of the hair under the specified lighting conditions is obtained. The lighting color parameters shift the hue of the hair color, the lighting intensity affects the overall brightness level, and the directionality determines the distribution range of the reflective area. The ambient reflection characteristics further adjust the color saturation to make the result closer to the realistic visual effect.

[0070] During illumination transformation, the color channels of each candidate hair color are individually modified. The red, green, and blue channels change to varying degrees according to the proportions of the spectral components of the illumination color, thus reproducing the overall modulation effect of the light source on the hair color. The transformed color value is the visual representation of the candidate hair color under that illumination scene, representing the visual effect of the hair color observed by the human eye in the corresponding environment. In this way, simulated representations of candidate hair colors can be generated under multiple typical illumination conditions, allowing subsequent adaptive evaluations to consider the visual stability of the hair color in different environments.

[0071] The overall adaptability index of each candidate hair color under all predefined lighting scene data is evaluated based on the visual performance color value; at least one recommended hair color is selected from the candidate hair colors according to the overall adaptability index; and the recommended hair color is output.

[0072] In another embodiment of the present invention, the comprehensive adaptability index for evaluating each candidate hair color includes:

[0073] For each candidate hair color, calculate the color difference between its visual representation color value under all predefined lighting scene data and the color difference value of the reference skin color feature. The color difference value is calculated in the color space using the color distance formula.

[0074] The color difference values ​​of all predefined lighting scene data are weighted and averaged. The weighted average result is used as a comprehensive adaptability index. The lower the comprehensive adaptability index value, the more stable the coordination between hair color and skin color under different lighting scenes. The weights are set based on the usage frequency data of the environment represented by each lighting scene data. The usage frequency data is obtained from user environment surveys.

[0075] It should be noted that when evaluating the comprehensive adaptability index of each candidate hair color, the calculation is performed by analyzing the color difference between the candidate hair color and the reference skin tone characteristics under different lighting conditions. For each candidate hair color, its visual color value after lighting transformation is obtained under all predefined lighting scene data, and these visual colors are compared with the reference skin tone characteristics under standard lighting conditions. The comparison process is carried out in a unified color space, which can reflect the comprehensive perceptual characteristics of human eyes for brightness, chromaticity, and color difference. The color difference value is calculated based on the distance relationship between color components to measure the visual proximity of two colors. When the visual color of the candidate hair color is close to the reference skin tone characteristics in space, it indicates that it has good coordination under that lighting condition. A color difference value is obtained for each lighting scene. In order to reflect the overall stability of the candidate hair color in different environments, the color difference values ​​of all lighting scenes are weighted and averaged. The weights are set according to the frequency of each lighting scene in users' daily lives. This frequency information is obtained through user environment survey data, which reflects the common proportions of users in natural light, indoor lighting, and nighttime lighting environments. The average result obtained by weighting the differences in each scene is used as a comprehensive adaptability index, which reflects the overall stability of hair color's harmony with skin tone under various lighting conditions. When the index value is low, it indicates that the hair color maintains good visual harmony in different environments and can be considered as one of the preferred hair colors.

[0076] It is worth noting that selecting at least one recommended hair color from the candidate hair colors includes:

[0077] Candidate hair colors are sorted according to the comprehensive adaptability index, from the lowest to the highest comprehensive adaptability index; the top-ranked candidate hair colors are selected as recommended hair colors, and the number of selections is a pre-set number.

[0078] A personalized recommendation system for a fully automatic hair dyeing and color matching machine includes:

[0079] Feature acquisition module: Acquire user images; extract user skin color features based on user images.

[0080] The user's skin tone features are processed for illumination correction to obtain reference skin tone features under standard lighting conditions.

[0081] Data processing module: Selects multiple candidate hair colors from the hair color database based on reference skin color features.

[0082] Acquire multiple predefined lighting scene data, each representing a typical ambient lighting condition; simulate and generate the visual representation color value of each candidate color based on the predefined lighting scene data.

[0083] Recommendation module: Evaluate the comprehensive adaptability index of each candidate hair color under all predefined lighting scene data based on the visual performance color value; select at least one recommended hair color from the candidate hair colors according to the comprehensive adaptability index; and output the recommended hair color.

[0084] The above-mentioned preset parameters and threshold selections shall be set by those skilled in the art according to the actual situation.

[0085] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A user-personalized recommendation method for a fully automatic hair dyeing and color matching machine, characterized in that, Includes the following steps: Acquire user image; extract user skin color features based on user image; The user's skin color features are subjected to illumination correction processing to obtain reference skin color features under standard illumination; Multiple candidate hair colors are selected from the hair color database based on the reference skin color features; Acquire multiple predefined lighting scene data, each of which represents a typical ambient lighting condition; The visual representation color value of each candidate hair color is generated based on predefined lighting scene data; The overall adaptability index of each candidate hair color under all predefined lighting scene data is evaluated based on the visual performance color value; at least one recommended hair color is selected from the candidate hair colors according to the overall adaptability index; and the recommended hair color is output.

2. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 1, characterized in that, Extracting user skin color features includes: Facial regions are identified from user images by analyzing contours and color distribution in the image; skin color pixel sets are located within facial regions, and if a pixel's color value belongs to a preset skin color range, then the pixel belongs to the skin color pixel set. The average color value is calculated as the user's skin color feature from the set of skin color pixels. The average color value is obtained by summing the color components of all skin color pixels and dividing by the number of pixels. The color components include red, green and blue components.

3. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 2, characterized in that, The illumination correction process includes: The illumination condition parameters of the user image are obtained by analyzing the highlight and shadow regions in the image. The highlight and shadow regions are determined based on the changes in pixel brightness values. The user's skin color features are adjusted to the color value under standard lighting conditions based on the lighting condition parameters. The standard lighting environment is defined as uniform white light conditions. The adjustment process is achieved through a color conversion function, which is built based on the lighting physical model. The lighting physical model takes into account the influence of lighting color temperature and intensity. The adjusted user skin tone features are used as reference skin tone features.

4. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 3, characterized in that, Selecting multiple candidate hair colors includes: Calculate the color similarity between the reference skin color feature and each hair color option in the hair color database. The color similarity is evaluated in a color space using Euclidean distance, and the color space is represented using RGB or LAB. The hair color options are sorted in descending order of color similarity, and the top-ranked hair colors are selected as candidate hair colors. The number of selections is based on a pre-set manual setting.

5. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 4, characterized in that, The visual representation color values ​​for each candidate hair color are generated based on predefined lighting scene data, including: Based on the lighting color and intensity parameters contained in the predefined lighting scene data, the original color values ​​of the candidate hair colors are subjected to lighting transformation. The lighting transformation is achieved through a color mixing process, which simulates the physical interaction between lighting and hair color, and superimposes the effect of the lighting color and intensity parameters onto the original color values ​​of the candidate hair colors. The color value output after the color mixing process is the visual representation color value of the candidate hair colors under the predefined lighting scene data.

6. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 5, characterized in that, The overall suitability metrics for evaluating each candidate hair color include: For each candidate hair color, calculate the color difference between its visual representation color value under all predefined lighting scene data and the color difference value of the reference skin color feature. The color difference value is calculated in the color space using the color distance formula. The color difference values ​​of all predefined lighting scene data are weighted and averaged. The weighted average result is used as a comprehensive adaptability index. The lower the comprehensive adaptability index value, the more stable the coordination between hair color and skin color under different lighting scenes. The weights are set based on the usage frequency data of the environment represented by each lighting scene data. The usage frequency data is obtained from user environment surveys.

7. The user-personalized recommendation method for a fully automatic hair dyeing and color matching machine according to claim 1, characterized in that, Select at least one recommended hair color from the candidate hair colors, including: Candidate hair colors are sorted according to the comprehensive adaptability index, from the lowest to the highest comprehensive adaptability index; the top-ranked candidate hair colors are selected as recommended hair colors, and the number of selections is a pre-set number.

8. A user-personalized recommendation system for a fully automatic hair dyeing and color matching machine, characterized in that, include: Feature acquisition module: Acquires user image; extracts user skin color features based on the user image; The user's skin color features are subjected to illumination correction processing to obtain reference skin color features under standard illumination; Data processing module: Selects multiple candidate hair colors from the hair color database based on the reference skin color features; Acquire multiple predefined lighting scene data, each of which represents a typical ambient lighting condition; The visual representation color value of each candidate hair color is generated based on predefined lighting scene data; Recommendation module: Evaluates the comprehensive adaptability index of each candidate hair color under all predefined lighting scene data based on the visual performance color value; selects at least one recommended hair color from the candidate hair colors according to the comprehensive adaptability index; and outputs the recommended hair color.

Images