Intelligent mirror virtual makeup intelligent recommendation system

Abstract

The application provides a smart mirror virtual makeup intelligent recommendation system, and relates to the technical field of data processing, and comprises an acquisition module configured to collect face image data of a user; based on the face image data, a plurality of feature regions of the face of the user are identified, including a left outer corner of the eye region, a right outer corner of the eye region and a lower nose tip region; center coordinates of the three feature regions are extracted, a face feature analysis circle frame is constructed, and based on the center, the radius and the distribution relationship of the three points relative to the center of the circle frame, structural characteristic analysis is performed on the face feature distribution in the frame to obtain face structure compensation parameters. The application realizes high matching of the virtual makeup effect with the actual scene and the skin state.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent mirror virtual makeup try-on intelligent recommendation system. Background Technology

[0002] With the accelerated digital transformation of beauty retail, virtual try-on technology has gradually matured based on computer vision, AR rendering and facial recognition technologies. Smart mirrors, as the core interactive terminal for offline stores and home scenarios, have been widely used in basic virtual try-on scenarios.

[0003] For example, a smart makeup mirror launched by a beauty brand collects static facial contours and skin tone data from a camera to virtually overlay preset makeup looks and recommend corresponding foundation and lipstick products based on skin tone brightness and color tone. However, it has obvious technical flaws. It only focuses on matching static facial features with fixed makeup templates and does not monitor dynamic physiological characteristics such as skin oil secretion and facial muscle movements in real time. It also does not integrate real-time environmental parameters such as ambient light intensity and color temperature. As a result, the virtual makeup effect is out of touch with the actual makeup application scenario and dynamic changes in the skin. The recommended product combinations are difficult to adapt to the makeup effect requirements and skin condition fluctuations in different environments. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide an intelligent mirror virtual makeup try-on intelligent recommendation system that achieves a high degree of consistency between the virtual makeup try-on effect and the actual scene and skin condition.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, a smart mirror virtual makeup try-on and intelligent recommendation system includes: The acquisition module is used to collect user facial image data; based on the facial image data, it identifies multiple feature regions of the user's face, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip; it extracts the center coordinates of the three feature regions, constructs a circular framework for facial feature analysis, and analyzes the structural characteristics of the facial feature distribution within the framework based on the distribution relationship of the center, radius, and three points relative to the center of the circle, thereby obtaining facial structure compensation parameters; The optimization module is used to optimize the initially identified static facial features through facial structure compensation parameters to obtain optimized static facial features; based on the optimized static facial features, the user's makeup requirements selected based on the static facial features are obtained through the touch screen of the smart mirror. The fusion module is used to obtain the current ambient light parameters through the ambient light sensor while obtaining the makeup requirements, and to fuse the static facial features, makeup requirements and ambient light parameters to obtain the initial makeup trial conditions. The monitoring module is used to monitor the dynamic physiological characteristics of the user's face in real time during the process of generating virtual makeup effects based on the initial makeup trial conditions. The matching module is used to match and analyze the dynamic physiological characteristics obtained from real-time monitoring with the preset cosmetic makeup-holding characteristic parameters to obtain the matching analysis results; The processing module is used to obtain virtual makeup effects that adapt to dynamic changes in the skin based on the initial makeup trial conditions and matching analysis results, and to recommend suitable product combinations.

[0006] Furthermore, user facial image data is collected; based on the facial image data, multiple feature regions of the user's face are identified, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip; the center coordinates of the three feature regions are extracted, and a circular framework for facial feature analysis is constructed. Based on the distribution relationship of the center, radius, and three points relative to the center of the circular framework, the structural characteristics of the facial feature distribution within the framework are analyzed to obtain facial structure compensation parameters, including: Receive the center coordinates of the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip region uploaded by the smart mirror, and obtain the set of facial feature point coordinates; The facial feature point coordinate set is used as input to perform geometric operations, construct a facial feature analysis circular frame that passes through three coordinate points, and obtain the center coordinates and radius value of the circular frame as the frame geometric parameters. Based on the frame geometric parameters, by calculating the distance and angle distribution of three feature points relative to the center of the circle, a set of structural relationship parameters describing the spatial layout of facial features is obtained; Using the set of structural relationship parameters as the basis for analysis, facial structure compensation parameters are obtained by quantitatively evaluating the uniformity, symmetry, and proportional relationship of feature distribution within the circular frame of facial feature analysis to correct the preliminary recognition results.

[0007] Furthermore, the initially identified facial static features are optimized using facial structure compensation parameters to obtain optimized facial static features. Based on these optimized features, the user's makeup preferences selected based on these features are obtained via the smart mirror's touchscreen, including: Based on facial structure compensation parameters, the spatial position and geometric proportion of the initially identified facial static features are corrected to obtain an optimized set of facial static feature parameters. The optimized set of static facial feature parameters is sent to the smart mirror terminal, and the facial feature analysis results are displayed to the user via the touch screen. The user receives makeup style selection instructions based on the facial feature analysis results displayed on the touchscreen, and obtains the corresponding makeup requirement parameters.

[0008] Furthermore, while acquiring makeup requirements, the system obtains current ambient light parameters through an ambient light sensor. It then fuses facial static features, makeup requirements, and ambient light parameters to obtain initial trial makeup conditions, including: By receiving the optimized set of facial static feature parameters and makeup requirement parameters, and simultaneously receiving the current ambient light parameters collected by the smart mirror ambient light sensor and transmitted via the network; The facial static feature parameter set, makeup requirement parameters and ambient light parameters are used as inputs, and a weighted fusion algorithm is used to obtain the environment-adaptive makeup rendering parameters. Based on makeup rendering parameters, structured initial makeup trial conditions are obtained by integrating facial features, makeup style, and lighting conditions.

[0009] Furthermore, during the process of generating virtual makeup effects based on initial makeup trial conditions, the dynamic physiological characteristics of the user's face are monitored in real time, including: While generating a virtual makeup trial effect based on the initial makeup trial conditions, the system initiates monitoring and analysis of the real-time video stream transmitted by the smart mirror. By performing time-series analysis on real-time video streams, dynamic change data of the user's face within a set time period can be extracted. Based on dynamically changing data, dynamic feature parameters reflecting the physiological state of the skin are identified through feature extraction algorithms; By integrating the dynamic feature parameters according to the time series, a set of user dynamic physiological feature parameters for makeup retention analysis is obtained.

[0010] Furthermore, the dynamic physiological characteristics obtained from real-time monitoring are matched with preset cosmetic makeup-holding characteristic parameters to obtain matching analysis results, including: By receiving the user's dynamic physiological characteristic parameter set and calling the preset cosmetic makeup-holding characteristic parameter library; The user's dynamic physiological characteristic parameter set is matched with various parameters in the cosmetic makeup retention characteristic parameter library in multiple dimensions to obtain a preliminary matching degree evaluation result; Based on the preliminary matching degree assessment results, a matching degree matrix between cosmetic characteristics and user physiological characteristics is constructed. The fitness matrix is ​​normalized and sorted to obtain the matching analysis results of recommendation priority ranking.

[0011] Furthermore, based on the initial makeup trial conditions and matching analysis results, a virtual makeup trial effect adapted to dynamic skin changes is obtained, and suitable product combinations are recommended, including: Based on the recommendation priority ranking in the matching analysis results, the makeup rendering parameters in the initial makeup trial conditions are dynamically adjusted to obtain dynamic makeup trial effect parameters for skin physiological changes. Based on the dynamic makeup effect parameters, the corresponding virtual makeup effect is obtained through rendering processing, and the makeup changes on the user's face over time are simulated. Based on the product matching relationships with the highest priority in the matching analysis results, the corresponding cosmetic information is extracted to obtain a suitable product combination recommendation list. The virtual makeup effect is then linked with the product combination recommendation list and output to the smart mirror terminal.

[0012] In a second aspect, a computing device includes: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to execute the system.

[0013] Thirdly, a computer-readable storage medium storing a program that, when executed by a processor, performs the system.

[0014] The above-described solution of the present invention has at least the following beneficial effects: Because it employs a technique that uses three points—the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip—to construct a circular framework for facial feature analysis and generate structural compensation parameters to optimize static facial features, it integrates static facial features, makeup requirements, and ambient light parameters to form initial makeup trial conditions. It also monitors dynamic facial physiological features in real time and performs multi-dimensional matching with preset cosmetic makeup-holding characteristic parameters, thereby dynamically adjusting makeup rendering parameters to generate trial makeup effects and recommended product combinations that adapt to skin changes. Therefore, it overcomes the technical problems of existing systems that rely solely on matching static facial features with fixed makeup templates, ignoring dynamic skin physiological changes and the influence of ambient light, resulting in distorted trial makeup effects and insufficient product recommendation adaptability. This achieves a high degree of consistency between virtual makeup trial effects and actual scenes and dynamic skin changes, making product recommendations more personalized and practical, and improving the user's makeup trial experience and decision-making efficiency. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of an intelligent mirror virtual makeup try-on intelligent recommendation system provided by an embodiment of the present invention.

[0016] Figure 2 This invention provides a flowchart illustrating the process of obtaining initial makeup trial conditions by acquiring current ambient light parameters through an ambient light sensor while simultaneously obtaining makeup requirements. The process involves fusing facial static features, makeup requirements, and ambient light parameters. Detailed Implementation

[0017] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0018] like Figure 1 As shown, an embodiment of the present invention proposes an intelligent mirror virtual makeup try-on intelligent recommendation system, including the following steps: The acquisition module is used to collect user facial image data; based on the facial image data, it identifies multiple feature regions of the user's face, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip; it extracts the center coordinates of the three feature regions, constructs a circular framework for facial feature analysis, and analyzes the structural characteristics of the facial feature distribution within the framework based on the distribution relationship of the center, radius, and three points relative to the center of the circle, thereby obtaining facial structure compensation parameters; The optimization module is used to optimize the initially identified static facial features through facial structure compensation parameters to obtain optimized static facial features; based on the optimized static facial features, the user's makeup requirements selected based on the static facial features are obtained through the touch screen of the smart mirror. The fusion module is used to obtain the current ambient light parameters through the ambient light sensor while obtaining the makeup requirements, and to fuse the static facial features, makeup requirements and ambient light parameters to obtain the initial makeup trial conditions. The monitoring module is used to monitor the dynamic physiological characteristics of the user's face in real time during the process of generating virtual makeup effects based on the initial makeup trial conditions. The matching module is used to match and analyze the dynamic physiological characteristics obtained from real-time monitoring with the preset cosmetic makeup-holding characteristic parameters to obtain the matching analysis results; The processing module is used to obtain virtual makeup effects that adapt to dynamic changes in the skin based on the initial makeup trial conditions and matching analysis results, and to recommend suitable product combinations.

[0019] In this embodiment of the invention, a facial feature analysis circular framework is constructed using three points—the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip—to generate facial structure compensation parameters and optimize static facial features. Simultaneously, these optimized static facial features, user makeup requirements, and ambient light parameters are integrated to form initial makeup trial conditions. During the generation of the virtual makeup trial effect, dynamic facial physiological features are monitored in real time and matched with preset cosmetic makeup-holding characteristic parameters in multiple dimensions. Ultimately, the makeup trial effect is dynamically adjusted, and suitable product combinations are recommended. Therefore, this overcomes the technical problems of existing smart mirror virtual makeup trial systems that rely solely on matching static facial features with fixed makeup templates, failing to consider dynamic skin physiological changes and the influence of ambient light, resulting in a disconnect between the makeup trial effect and the actual scene, and insufficient product recommendation adaptability. This achieves the technical effect of accurately adapting the virtual makeup trial effect to dynamic skin changes and the actual environment, making product recommendations more personalized and practical, and improving the user's makeup trial experience and purchase decision efficiency.

[0020] In a preferred embodiment of the present invention, facial image data of the user is acquired; based on the facial image data, multiple feature regions of the user's face are identified, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip; the center coordinates of the three feature regions are extracted, and a circular framework for facial feature analysis is constructed. Based on the distribution relationship of the center, radius, and three points relative to the center of the circular framework, the structural characteristics of the facial feature distribution within the framework are analyzed to obtain facial structure compensation parameters, including: The system receives the center coordinates of the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip from the smart mirror, and obtains a set of facial feature point coordinates. Specifically, the smart mirror takes a picture of the user's face through its built-in image acquisition device, accurately locates the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip, determines the center position of each of these three feature areas and records the corresponding coordinate information, then transmits the coordinate information, and after receiving the coordinate information, it classifies and organizes it to finally form a complete set of facial feature point coordinates.

[0021] Using a set of facial feature point coordinates as input, geometric operations are performed to construct a circular framework for facial feature analysis that passes through three coordinate points. The coordinates of the center and the radius of the circular framework are then obtained as the geometric parameters of the framework. Specifically, the following steps are taken: the prepared set of facial feature point coordinates is used as the core data input, a dedicated geometric analysis process is initiated, and a series of geometric operations are performed to determine a circular framework for facial feature analysis that can simultaneously pass through the center coordinates of the left outer corner of the eye, the center coordinates of the right outer corner of the eye, and the center coordinates of the lower part of the nose tip. During the operation, the coordinate values ​​corresponding to the specific position of the center of the circular framework and the radius of the circular framework are accurately calculated. The calculated center coordinates and radius values ​​are then uniformly recorded as the geometric parameters of the framework and saved.

[0022] Based on the frame geometry parameters, a set of structural relationship parameters describing the spatial layout of facial features is obtained by calculating the distance and angle distribution of three feature points relative to the center of the circle. Specifically, this includes: retrieving the saved frame geometry parameters, calculating the straight-line distance from each feature point to the center of the circle for the three feature points: the center of the outer corner of the left eye, the center of the outer corner of the right eye, and the center of the lower part of the nose tip; and simultaneously measuring the angular position of each feature point relative to the center of the circle. All the calculated distance data and angle distribution data are then summarized and organized to form a set of structural relationship parameters that comprehensively describes the spatial layout of facial features.

[0023] Using the structural relationship parameter set as the basis for analysis, the uniformity, symmetry, and proportional relationship of the feature distribution within the circular frame of facial feature analysis are quantitatively evaluated to obtain facial structure compensation parameters for correcting the preliminary recognition results. Specifically, this includes: using the structural relationship parameter set as the core basis for analysis, initiating a quantitative evaluation program to conduct comprehensive numerical analysis and objective evaluation of the uniformity of facial feature distribution within the circular frame, the symmetry between different features, and the proportional relationship between each feature; identifying deviations in the spatial position and geometric proportion of the initially recognized facial features based on the evaluation results; and generating facial structure compensation parameters for correcting the preliminary recognition results based on the recognition deviations.

[0024] In this embodiment of the invention, a feature point coordinate set is formed by receiving the center coordinates of three points: the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip. A facial feature analysis circular frame passing through these three points is constructed through geometric operations, and geometric parameters such as the center and radius of the circle are obtained. Then, the distance and angle distribution of the three points relative to the center of the circle are calculated to obtain a structural relationship parameter set. Finally, based on this parameter set, the uniformity, symmetry, and proportional relationship of the feature distribution within the frame are quantitatively evaluated to generate facial structure compensation parameters. Therefore, this invention overcomes the technical problems of spatial position deviation and inaccurate geometric proportion in the initial recognition results of traditional facial feature recognition, thereby achieving accurate correction of the initial recognition error of facial static features and improving the accuracy and reliability of facial feature analysis.

[0025] In a preferred embodiment of the present invention, facial structure compensation parameters are used to optimize the initially identified facial static features to obtain optimized facial static features. Based on the optimized facial static features, the user's makeup preferences selected based on the facial static features are obtained through the touchscreen of the smart mirror, including: Based on facial structure compensation parameters, the spatial position and geometric proportion of the initially identified facial static features are corrected to obtain an optimized set of facial static feature parameters. Specifically, this includes: calling the previously generated facial structure compensation parameters, processing the initially identified facial static features based on the parameters, focusing on adjusting the spatial distribution of facial static features, correcting geometric proportion deviations between features, and accurately correcting each facial static feature one by one, finally integrating them to form an optimized set of facial static feature parameters with accurate data.

[0026] The optimized set of static facial feature parameters is sent to the smart mirror terminal, and the facial feature analysis results are displayed to the user through the touch screen. Specifically, the optimized set of static facial feature parameters is sent to the smart mirror terminal through the data transmission channel. After receiving the parameter set, the smart mirror terminal parses and processes it, transforming the key information in the parameter set into facial feature analysis results that are easy for the user to understand. Then, it displays the results to the user in an intuitive form through its own touch screen, allowing the user to clearly understand the specific situation of their own facial features.

[0027] The system uses the facial feature analysis results displayed on the touchscreen as input to select a makeup style, thereby obtaining the corresponding makeup requirement parameters. Specifically, the smart mirror touchscreen provides multiple makeup style options for users to choose from while displaying the facial feature analysis results. Users can input makeup style selection commands that meet their needs by clicking or swiping on the touchscreen, based on the displayed facial feature analysis results. The system receives and analyzes these selection commands in real time, ultimately converting them into the corresponding makeup requirement parameters.

[0028] In this embodiment of the invention, because a technical means is adopted to correct the spatial position and geometric proportion of the initially identified facial static features based on facial structure compensation parameters, send the optimized set of facial static feature parameters to the smart mirror terminal and display the analysis results through the touch screen, and then receive the user's makeup style selection command based on the analysis results through the touch screen, the technical problem of spatial position and geometric proportion deviation of the initially identified facial static features and the difficulty for users to select a suitable makeup style based on their own accurate facial features is overcome. Thus, the accuracy of facial static feature parameters is improved, allowing users to clearly understand their own facial features and make makeup choices that fit their personal conditions.

[0029] like Figure 2 As shown, in another preferred embodiment of the present invention, while obtaining the makeup requirements, the current ambient light parameters are obtained through an ambient light sensor. The facial static features, makeup requirements, and ambient light parameters are then fused to obtain the initial makeup trial conditions, including: By receiving the optimized set of facial static feature parameters and makeup requirement parameters, and simultaneously receiving the current ambient light parameters collected by the smart mirror's ambient light sensor and transmitted via the network, the process includes: first, receiving the optimized set of facial static feature parameters from the previous steps; second, receiving the makeup requirement parameters input and parsed by the user through the smart mirror's touchscreen; and third, simultaneously receiving the current ambient light-related data collected in real time by the smart mirror's built-in ambient light sensor. The ambient light data is transmitted via the network, and the three types of parameters are received and initially processed to ensure that the data is complete and effective in real time.

[0030] Using facial static feature parameters, makeup requirement parameters, and ambient light parameters as input, a weighted fusion algorithm is used to obtain environment-adapted makeup rendering parameters. Specifically, this involves classifying and decomposing the input facial static feature parameters, makeup requirement parameters, and ambient light parameters, clarifying the specific indicators included in each category. Facial static feature parameters include skin tone brightness, skin tone hue, facial contour proportions, and relative positions of facial features; makeup requirement parameters include makeup style, key areas to be emphasized, and the intensity of makeup; ambient light parameters include the current ambient light intensity, color temperature, and light uniformity. Based on professional knowledge in the beauty industry and historical makeup trial feedback data, a parameter influence weight evaluation model is constructed, assigning initial weights to each category of parameters and its specific indicators. For example, the color temperature in ambient light significantly affects the color rendering of the foundation, so its weight is set relatively high; the user-specified makeup style plays a dominant role in the makeup requirement parameters, with a higher weight than other sub-indicators; and skin tone brightness in facial static features is directly related to foundation color matching, with a higher weight than auxiliary features such as contour proportions.

[0031] All parameters are standardized, transforming parameters of different magnitudes and dimensions into a unified range of 0-1 to eliminate fusion bias caused by differences in dimensions and ensure that all parameters are equally comparable in calculation. Then, parameter fusion is achieved through weighted calculation, multiplying each standardized parameter by its corresponding weight to obtain the weighted value of a single indicator. The weighted values ​​of all indicators under the same parameter category are summed to obtain the comprehensive weighted value of three categories of parameters: facial static features, makeup requirements, and ambient light. Finally, based on the category weights of the three types of parameters, the comprehensive weighted values ​​of the three categories are summed again to generate a preliminary makeup rendering parameter set, including specific rendering indicators such as base makeup transparency, eyeshadow saturation, lipstick gloss, and blush color intensity. At the same time, the algorithm has a built-in dynamic weight adjustment mechanism, receiving user satisfaction data on the makeup trial effect from the smart mirror in real time. If it is found that the actual impact of a certain parameter deviates significantly from the initial weight, the weight value of the corresponding parameter is automatically corrected so that the final output makeup rendering parameters accurately match the user's facial features and makeup requirements.

[0032] Based on makeup rendering parameters, structured initial makeup trial conditions are obtained by integrating facial features, makeup style, and lighting conditions. Specifically, based on the generated environment-adaptive makeup rendering parameters, the user's facial feature information, selected makeup style information, and current lighting condition information are fully integrated. This information is then classified, sorted, and organized according to a preset structured format to form complete initial makeup trial conditions that include three core dimensions: facial features, makeup style, and lighting conditions.

[0033] In this embodiment of the invention, because it adopts the technical means of receiving the optimized set of facial static feature parameters, makeup requirement parameters and the current ambient light parameters collected by the smart mirror ambient light sensor, and obtaining the environment-adapted makeup rendering parameters through a weighted fusion algorithm, and then integrating them to form a structured initial makeup trial condition that includes facial features, makeup style and lighting conditions, it overcomes the technical problem of existing systems that do not integrate ambient light parameters and rely only on facial static features and fixed makeup templates, resulting in the makeup trial condition being out of sync with the actual lighting environment. Thus, it achieves the goal of enabling the initial makeup trial condition to accurately adapt to the user's facial features, makeup requirements and current ambient light conditions.

[0034] In a preferred embodiment of the present invention, during the process of generating a virtual makeup effect based on initial makeup trial conditions, the dynamic physiological characteristics of the user's face are monitored in real time, including: While generating the virtual makeup effect based on the initial makeup trial conditions, the system simultaneously initiates monitoring and analysis of the real-time video stream transmitted by the smart mirror. Specifically, based on the previously established structured initial makeup trial conditions, the system starts the virtual makeup effect generation program and begins rendering the corresponding virtual makeup scene. At the same time, the system activates a dedicated video stream monitoring module to receive the user's facial video stream, which is captured and transmitted in real-time by the smart mirror's camera, and continuously monitors and analyzes this video stream.

[0035] By performing time-series analysis on real-time video streams, dynamic change data of the user's face within a set time period is extracted. Specifically, this includes: analyzing and comparing the received real-time video stream frame by frame in chronological order, conducting time-series analysis, and continuously tracking subtle changes in the user's face within a pre-set time period, such as changes in skin gloss, contraction and relaxation of facial muscles, and color depth fluctuations in specific areas, thereby extracting raw data reflecting dynamic changes in the face.

[0036] Based on dynamically changing data, a feature extraction algorithm is used to identify dynamic feature parameters reflecting the physiological state of the skin. Specifically, this involves running a feature extraction algorithm based on extracted facial dynamic change data. The algorithm filters and analyzes the data to identify key dynamic features that directly reflect the physiological state of the skin, such as the reflectivity parameters caused by sebum secretion, the amplitude and frequency parameters of facial muscle movements, and the texture parameters corresponding to changes in skin humidity, thus forming specific dynamic feature parameters.

[0037] By integrating dynamic feature parameters according to a time series, a set of user dynamic physiological feature parameters for makeup-holding characteristic analysis is obtained. Specifically, this includes arranging and integrating the identified dynamic feature parameters according to their chronological order of appearance on the time axis, and recording the parameter values ​​and trends at each time point. In this way, a complete dataset reflecting the changes in the user's facial physiological state over time is constructed, which is the user dynamic physiological feature parameter set used for makeup-holding characteristic analysis.

[0038] In this embodiment of the invention, because it employs a technical approach that simultaneously generates a virtual makeup effect based on initial makeup conditions, monitors and analyzes the real-time video stream transmitted by the smart mirror, extracts facial dynamic change data within a set time period through time-series analysis, identifies dynamic feature parameters reflecting the skin's physiological state using feature extraction algorithms, and finally integrates these parameters into a set of user dynamic physiological feature parameters according to the time sequence, it overcomes the existing technical problem of not monitoring the user's skin oil secretion, facial muscle movement, and other dynamic physiological characteristics in real time, leading to a disconnect between the virtual makeup effect and the skin's dynamic changes. This achieves the technical effect of accurately capturing the user's real-time facial physiological state changes and providing reliable dynamic data support for generating a makeup effect that adapts to the skin's dynamic changes.

[0039] In a preferred embodiment of the present invention, the dynamic physiological characteristics obtained through real-time monitoring are matched and analyzed with preset cosmetic makeup-holding characteristic parameters to obtain matching analysis results, including: By receiving a set of dynamic physiological characteristics parameters of the user and calling a pre-set cosmetic makeup-holding characteristic parameter library, the process includes: receiving the set of dynamic physiological characteristics parameters of the user integrated in the previous steps, which contains data reflecting the changes in the user's physiological state over time, such as skin oil secretion, facial expressions, and muscle movements; and simultaneously starting a parameter library calling program to retrieve the pre-set cosmetic makeup-holding characteristic parameter library, which stores relevant characteristic parameters of various cosmetics, such as oil resistance, friction resistance, and makeup-holding time.

[0040] The user's dynamic physiological characteristic parameter set is matched with various parameters in the cosmetic makeup-holding characteristic parameter library in multiple dimensions to obtain a preliminary matching degree assessment result. Specifically, this includes: comparing and calculating various parameters in the user's dynamic physiological characteristic parameter set with the corresponding characteristic parameters of each cosmetic in the cosmetic makeup-holding characteristic parameter library in multiple dimensions; the comparison dimensions include the degree of matching between the user's skin oil secretion and the cosmetic's oil control characteristics, the compatibility between the user's facial muscle movement amplitude and the cosmetic's anti-friction characteristics, and the fit between the user's skin dynamic change rate and the cosmetic's makeup-holding time, etc., to obtain a preliminary matching degree assessment result through comprehensive calculation and analysis.

[0041] Based on the preliminary matching degree assessment results, an adaptation matrix between cosmetic characteristics and user physiological characteristics is constructed. Specifically, this includes: based on the preliminary matching degree assessment results, a dedicated adaptation matrix is ​​constructed. The row dimension of the matrix corresponds to the makeup-holding characteristic parameter categories of various cosmetics, and the column dimension corresponds to different parameter dimensions of the user's dynamic physiological characteristics. Each cell in the matrix is ​​filled with the corresponding matching degree assessment value, clearly presenting the adaptation relationship between cosmetic characteristics and user physiological characteristics.

[0042] The fit matrix is ​​normalized and sorted to obtain matching analysis results with recommendation priority ranking. Specifically, the following steps are taken: the constructed fit matrix is ​​normalized to convert the matching degree values ​​of different ranges in the matrix to the same data interval, eliminating the comparison bias caused by the difference in numerical range; then, the normalized fit matrix is ​​sorted and analyzed, and the various cosmetics are arranged in descending order of matching degree value, finally obtaining matching analysis results with recommendation priority ranking.

[0043] In this embodiment of the invention, by employing a technical approach that receives a set of dynamic physiological characteristic parameters of the user and calls a pre-set database of cosmetic makeup-holding characteristic parameters, performs multi-dimensional matching calculations on the two types of parameters to obtain a preliminary matching degree evaluation result, and then constructs an adaptation degree matrix based on the result and performs normalization and ranking analysis, the technical problem of existing technologies failing to associate the user's dynamic skin physiological characteristics with the makeup-holding characteristics of cosmetics, resulting in the difficulty of recommending product combinations to adapt to skin condition fluctuations, is overcome. This achieves the goal of accurately mining the adaptation relationship between the user's dynamic skin condition and the makeup-holding characteristics of cosmetics, and obtaining matching results with recommendation priority.

[0044] In a preferred embodiment of the present invention, based on initial makeup trial conditions and matching analysis results, a virtual makeup trial effect adapted to dynamic changes in skin is obtained, and suitable product combinations are recommended, including: Based on the recommendation priority ranking in the matching analysis results, the makeup rendering parameters in the initial makeup try-on conditions are dynamically adjusted to obtain dynamic makeup try-on effect parameters that reflect changes in skin physiology. Specifically, this includes: obtaining the previously obtained matching analysis results with recommendation priority ranking; using the ranking results as the core basis, identifying the makeup-holding characteristic parameters corresponding to the higher-priority cosmetics; and combining these parameters to make targeted dynamic adjustments to the existing makeup rendering parameters in the initial makeup try-on conditions. For example, adjusting the gloss parameter of the base makeup based on the oil-control characteristics of high-priority products, and adjusting the adhesion parameter of the makeup based on its anti-friction characteristics, ultimately obtaining dynamic makeup try-on effect parameters that adapt to changes in skin physiology.

[0045] Based on the dynamic makeup try-on effect parameters, the corresponding virtual makeup try-on effect is obtained through rendering processing, and the makeup's changes over time on the user's face are simulated. Specifically, this includes: inputting the dynamic makeup try-on effect parameters into the AR rendering program, and generating a virtual makeup try-on screen that highly matches the user's skin condition and environmental conditions through rendering processing. At the same time, the makeup change simulation program is launched, simulating the changes in the makeup's state over different time periods based on the changing trends of the user's dynamic physiological characteristic parameters. For example, the slight makeup fading effect caused by increased sebum secretion over time, and the natural smudging process of eye makeup edges caused by facial muscle movements, completely recreating the entire makeup's lasting process in a real-world scenario.

[0046] Based on the highest-priority product matching relationships in the matching analysis results, the corresponding cosmetic information is extracted to obtain a suitable product combination recommendation list. The virtual makeup effect is then linked with the product combination recommendation list and output to the smart mirror terminal. Specifically, this involves: filtering out the highest-priority product matching relationships from the matching analysis results, extracting detailed product information, including product name, shade, core ingredients, makeup lasting advantages, and suitable skin conditions, and integrating this information to form a clearly structured suitable product combination recommendation list; subsequently, the generated virtual makeup effect is linked and bound to the product combination recommendation list and sent to the smart mirror terminal via a data transmission channel. Upon receiving the data, the smart mirror terminal simultaneously displays the virtual makeup effect and corresponding product recommendation information on the touchscreen, allowing users to intuitively view the makeup effect and understand the details of the recommended products.

[0047] In this embodiment of the invention, by employing a technique of adjusting the makeup rendering parameters in the initial makeup trial conditions based on the recommendation priority ranking of matching analysis results, rendering the virtual makeup trial effect according to the dynamic makeup trial effect parameters and simulating the makeup wear change process, extracting the product information with the highest priority to form a recommended list of suitable product combinations and outputting it to the smart mirror terminal in association with the makeup trial effect, the technical problem of existing systems' virtual makeup trial effect being disconnected from the actual makeup wear scenario and dynamic changes of the skin, and the recommended product combinations being difficult to adapt to skin condition fluctuations, is overcome. This achieves the technical effect that the generated virtual makeup trial effect can accurately adapt to dynamic changes of the skin and realistically reproduce the makeup wear process, making product recommendations more targeted and adaptable, and improving the user's makeup trial experience and purchase decision efficiency.

[0048] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0049] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the system as described above. All implementations in the above system embodiments are applicable to this embodiment and can achieve the same technical effects.

[0050] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A smart mirror virtual makeup try-on intelligent recommendation system, characterized in that, include: The acquisition module is used to collect user facial image data; Based on facial image data, multiple feature regions of the user's face are identified, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip. The center coordinates of the three feature regions are extracted, and a circular framework for facial feature analysis is constructed. Based on the distribution relationship of the center, radius, and three points relative to the center of the circular framework, the structural characteristics of the facial feature distribution within the framework are analyzed to obtain facial structure compensation parameters. The optimization module is used to optimize the initially identified static facial features using facial structure compensation parameters to obtain optimized static facial features. Based on optimized facial static features, the smart mirror's touchscreen obtains the user's makeup preferences selected based on these features. The fusion module is used to obtain the current ambient light parameters through the ambient light sensor while obtaining the makeup requirements, and to fuse the static facial features, makeup requirements and ambient light parameters to obtain the initial makeup trial conditions. The monitoring module is used to monitor the dynamic physiological characteristics of the user's face in real time during the process of generating virtual makeup effects based on the initial makeup trial conditions. The matching module is used to match and analyze the dynamic physiological characteristics obtained from real-time monitoring with the preset cosmetic makeup-holding characteristic parameters to obtain the matching analysis results; The processing module is used to obtain virtual makeup effects that adapt to dynamic changes in the skin based on the initial makeup trial conditions and matching analysis results, and to recommend suitable product combinations.

2. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 1, characterized in that, Collect user facial image data; based on the facial image data, identify multiple feature regions of the user's face, including the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip; extract the center coordinates of the three feature regions, construct a circular framework for facial feature analysis, and based on the distribution relationship of the center, radius, and three points relative to the center of the circular framework, perform structural characteristic analysis on the distribution of facial features within the framework to obtain facial structure compensation parameters, including: Receive the center coordinates of the outer corner of the left eye, the outer corner of the right eye, and the lower part of the nose tip region uploaded by the smart mirror, and obtain the set of facial feature point coordinates; The facial feature point coordinate set is used as input to perform geometric operations, construct a facial feature analysis circular frame that passes through three coordinate points, and obtain the center coordinates and radius value of the circular frame as the frame geometric parameters. Based on the frame geometric parameters, by calculating the distance and angle distribution of three feature points relative to the center of the circle, a set of structural relationship parameters describing the spatial layout of facial features is obtained; Using the set of structural relationship parameters as the basis for analysis, facial structure compensation parameters are obtained by quantitatively evaluating the uniformity, symmetry, and proportional relationship of feature distribution within the circular frame of facial feature analysis to correct the preliminary recognition results.

3. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 2, characterized in that, The initially identified static facial features are optimized by using facial structure compensation parameters to obtain optimized static facial features. Based on optimized facial static features, the smart mirror's touchscreen retrieves the user's makeup preferences selected based on these features, including: Based on facial structure compensation parameters, the spatial position and geometric proportion of the initially identified facial static features are corrected to obtain an optimized set of facial static feature parameters. The optimized set of static facial feature parameters is sent to the smart mirror terminal, and the facial feature analysis results are displayed to the user via the touch screen. The user receives makeup style selection instructions based on the facial feature analysis results displayed on the touchscreen, and obtains the corresponding makeup requirement parameters.

4. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 3, characterized in that, While acquiring makeup requirements, the system uses an ambient light sensor to obtain current ambient light parameters. It then fuses facial static features, makeup requirements, and ambient light parameters to obtain initial trial makeup conditions, including: By receiving the optimized set of facial static feature parameters and makeup requirement parameters, and simultaneously receiving the current ambient light parameters collected by the smart mirror ambient light sensor and transmitted via the network; The facial static feature parameter set, makeup requirement parameters and ambient light parameters are used as inputs, and a weighted fusion algorithm is used to obtain the environment-adaptive makeup rendering parameters. Based on makeup rendering parameters, structured initial makeup trial conditions are obtained by integrating facial features, makeup style, and lighting conditions.

5. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 4, characterized in that, During the process of generating virtual makeup effects based on initial makeup trial conditions, the dynamic physiological characteristics of the user's face are monitored in real time, including: While generating a virtual makeup trial effect based on the initial makeup trial conditions, the system initiates monitoring and analysis of the real-time video stream transmitted by the smart mirror. By performing time-series analysis on real-time video streams, dynamic change data of the user's face within a set time period can be extracted. Based on dynamically changing data, dynamic feature parameters reflecting the physiological state of the skin are identified through feature extraction algorithms; By integrating the dynamic feature parameters according to the time series, a set of user dynamic physiological feature parameters for makeup retention analysis is obtained.

6. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 5, characterized in that, The dynamic physiological characteristics obtained from real-time monitoring are matched and analyzed with preset cosmetic makeup-holding properties parameters to obtain the matching analysis results, including: By receiving the user's dynamic physiological characteristic parameter set and calling the preset cosmetic makeup-holding characteristic parameter library; The user's dynamic physiological characteristic parameter set is matched with various parameters in the cosmetic makeup retention characteristic parameter library in multiple dimensions to obtain a preliminary matching degree evaluation result; Based on the preliminary matching degree assessment results, a matching degree matrix between cosmetic characteristics and user physiological characteristics is constructed. The fitness matrix is ​​normalized and sorted to obtain the matching analysis results of recommendation priority ranking.

7. The intelligent mirror virtual makeup try-on intelligent recommendation system according to claim 6, characterized in that, Based on the initial makeup trial conditions and matching analysis results, a virtual makeup trial effect that adapts to dynamic changes in skin is obtained, and suitable product combinations are recommended, including: Based on the recommendation priority ranking in the matching analysis results, the makeup rendering parameters in the initial makeup trial conditions are dynamically adjusted to obtain dynamic makeup trial effect parameters for skin physiological changes. Based on the dynamic makeup effect parameters, the corresponding virtual makeup effect is obtained through rendering processing, and the makeup changes on the user's face over time are simulated. Based on the product matching relationships with the highest priority in the matching analysis results, the corresponding cosmetic information is extracted to obtain a suitable product combination recommendation list. The virtual makeup effect is then linked with the product combination recommendation list and output to the smart mirror terminal.

8. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to perform the system as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, performs the system as described in any one of claims 1 to 7.

Images