Light source adjusting method and device, intelligent beauty makeup method and beauty makeup mirror
By calculating the color difference value of the object to be identified for light source adjustment, the problem that existing beauty mirror light source adjustment relies on complex spectral recognition is solved, and efficient and accurate light source adjustment is achieved, which reduces hardware cost and volume and improves user experience.
Patent Information
- Application Number
- CN202410828067.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-25
- Publication Date
- 2024-11-01
AI Technical Summary
Existing beauty mirrors rely on complex spectral recognition technology in light source adjustment, with high hardware requirements and complex calculations, resulting in increased equipment size, high cost and poor user experience, and cannot provide personalized light source adjustment, affecting the accuracy of makeup effect. sex.
By obtaining the real-time color parameters and preset color parameters of the object to be identified, the color difference value is calculated, and the light source is adjusted based on this, until the difference value is reduced to the preset range, reducing the calculation amount, improving the adjustment efficiency, and adapting to the light of the human eye. The actual response reduces hardware demand and reduces the cost of beauty mirrors.
It achieves the efficiency of light source adjustment, improves the accuracy of makeup effects, reduces the hardware cost and volume of beauty mirrors, enhances its portability and practicality, provides makeup simulation closer to natural light, and improves user experience.
Smart Images

Figure CN118890746A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent beauty technology, and in particular to a light source adjustment method and device, an intelligent beauty method and a beauty mirror. Background Art
[0002] In the prior art, the traditional makeup mirror has a simple structure, mainly with the function of reflecting the mirror surface, and is only used as an auxiliary tool for makeup. It cannot provide in-depth skin analysis and personalized makeup guidance. Users need to rely on their own experience or refer to external materials to choose skin care products and makeup techniques, which lacks pertinence and accuracy. At the same time, although there are some beauty apps or software on the market that can provide makeup tutorials and videos, these tools are usually separated from the makeup mirror, and users need to switch back and forth between the mirror and the mobile phone or computer, which is inconvenient to operate; in addition, skin care products and beauty tools are becoming more and more abundant, and repeated purchases and poor management of the shelf life and cleanliness of the tools have a negative impact on skin management.
[0003] Existing technologies have obvious deficiencies in providing in-depth skin analysis, personalized makeup guidance, simulating different lighting conditions, and integrating makeup tutorials and videos. These problems limit users' access to better and more convenient skin care and makeup experiences.
[0004] In the related art, spectral recognition is usually used to adjust the light source. As an advanced optical analysis method, spectral recognition technology can accurately detect and identify the wavelength distribution and intensity of different light sources. By identifying the ambient light of the beauty mirror and adjusting the light source of the beauty mirror, the beauty mirror can simulate a variety of natural light and artificial light environments, thereby helping users to better preview the performance of makeup under different light conditions. However, spectral recognition is highly dependent on high-precision hardware design, such as precision spectrometers, high-quality optical lenses, and sensitive photodetectors. These hardware together ensure the accuracy and reliability of spectral data. Similarly, spectral recognition also involves complex software program development. The software needs to be able to process and analyze the spectral data returned from the hardware in real time, and dynamically adjust the light source output of the beauty mirror based on this data. This requires not only that the software has powerful data processing capabilities, but also that there must be a sophisticated control algorithm to ensure smooth and accurate light source adjustment. The hardware installation of spectral recognition will not only increase the size of the beauty mirror, but also increase the manufacturing cost of the beauty mirror, which is not conducive to widespread application. Ordinary light source adjustment is limited to brightening or dimming the light source, and adjusting the brightness depends on the user's subjective consciousness and visual perception. Since subjective adjustment ignores the influence of factors such as ambient light and object reflectivity on the light source effect, the adjustment effect of the light source is prone to deviation, which affects the user's makeup effect, causing the user to deviate from the judgment of their own makeup and affecting the user's experience.
[0005] Chinese patent "A smart makeup method and smart makeup mirror", publication number: CN 109671142A, publication date: April 23, 2019, specifically discloses a makeup method including: collecting the current video frame; detecting whether there is a face in the current video frame and returning feature point location information; generating a 3D model consistent with the current user's facial features based on the feature point location information; using parallel light sources for orthogonal projection, so that the 3D model generates a projection at the imaging position of the video frame, generating visual difference, so that when observed from the OpenGL space, the 3D model completely blocks the face of the video image frame; using OpenGL Shader to render the makeup product to the corresponding feature point position of the 3D model, and setting the 3D model rendering effect to transparent; calculating the distance between the user and the OpenGL observation point based on the distance between the two specific points of the face, calculating the 3D model scaling ratio based on this distance, and deflecting the 3D model in real time based on the spatial deflection angle information returned by face tracking, generating a real-time makeup trial effect under face movements. This solution renders according to the makeup product, that is, using universal makeup techniques or trendy makeup, and has poor adaptability.
[0006] The Chinese patent "Bathroom mirror and its control method and control device", publication number: CN 109315993 A, publication date: February 12, 2019, specifically discloses that the above control method includes the following steps: obtaining a panoramic image of cosmetics; identifying the panoramic image of cosmetics to obtain relevant information of cosmetics, wherein the relevant information of cosmetics includes an expiration date; obtaining the current date to calculate the difference between the expiration date and the current date, and judging whether the difference between the expiration date and the current date is less than a preset threshold; if the difference between the expiration date and the current date is less than the preset threshold, generating a reminder message that the cosmetics are about to expire; and providing the reminder message that the cosmetics are about to expire to the user through the bathroom mirror. This solution only provides the reminder message of the expiration to the user, and cannot provide the user with adaptation suggestions. It needs to rely on the user's own consideration, and there will be long-term reminders before the cosmetics expire, which has a poor user experience. Summary of the invention
[0007] The present application aims at the problem that light source adjustment depends on spectral recognition, has complex calculations and high hardware requirements in the prior art, and provides a light source adjustment method and device, an intelligent beauty makeup method and a beauty mirror. According to the real-time color parameters and preset color parameters of the object to be identified, the difference between the color of the object to be identified under the current light source and its actual color is obtained, and a color difference value is obtained based on the color difference and the light source adjustment is performed on this basis until the color difference value is reduced to a preset range, so that the color of the object to be identified under the light source condition is close to the actual color, and there is no need to perform spectral recognition and comparison of the light source light, which reduces the amount of calculation, improves the adjustment efficiency, and can adapt to the actual reaction of the human eye to light, thereby improving the accuracy of the user's simulated makeup effect generation and improving the user experience. In the present application, the color difference value calculation based on image processing does not require additional hardware design, reduces the hardware loading on the beauty mirror, reduces the manufacturing requirements of the beauty mirror, and meets the light source adjustment requirements, reduces hardware costs, improves the portability of the beauty mirror, and thus improves the practicality of the beauty mirror, which is easy to use.
[0008] In order to achieve the above technical objectives, the present application provides a technical solution, which is a light source adjustment method, comprising the following steps: obtaining an object to be identified, confirming preset color parameters of the object to be identified, and identifying real-time color parameters of the object to be identified; obtaining a color difference value based on the difference between the real-time color parameters and the preset color parameters; and performing light source adjustment according to the color difference value until the color difference value is reduced to a preset range. According to the real-time color parameters and preset color parameters of the object to be identified, the difference between the color of the object to be identified under the current light source and its actual color is obtained, and the color difference value is obtained based on the color difference and the light source adjustment is performed until the color difference value is reduced to a preset range, so that the color of the object to be identified under the light source conditions is close to the actual color, and there is no need to perform spectral identification and comparison of the light source light, thereby reducing the amount of calculation, improving the adjustment efficiency, and being able to adapt to the actual reaction of the human eye to light.
[0009] Furthermore, the method of obtaining a color difference value based on the difference between the real-time color parameters and the preset color parameters also includes: screening the real-time color parameters to obtain real-time background color parameters and real-time foreground color parameters; performing color compensation according to the real-time background color parameters and the real-time foreground color parameters to obtain updated color parameters; and obtaining a color difference value by the difference between the updated color parameters and the preset color parameters.
[0010] Furthermore, it also includes: based on historical training pictures, obtaining background color parameters located at the transition boundary in the historical training pictures, as well as first foreground color parameters and second foreground color parameters located at the center point; according to the matching of the first foreground color parameters and the second foreground color parameters, obtaining color compensation values corresponding to the background color parameters, and constructing a color compensation model with the background color parameters, the first color parameters and the color compensation values.
[0011] Another technical solution provided by the present application is an intelligent makeup method, comprising the following steps: obtaining user real-time data, and performing the above-mentioned light source adjustment based on the user real-time data; obtaining user update data after light source adjustment, and constructing a user face three-dimensional model according to the user update data; and generating a simulated makeup effect with the user face three-dimensional model and preset makeup data. Since the recognition and simulation of user makeup depends on the accurate recognition of color, in order to improve the color rendering of makeup, it is necessary to adjust the light source so that the recognized and simulated makeup effects are closer to the actual light source, so as to facilitate the user's accurate perception of their own makeup. Therefore, the user real-time data is obtained, and the light source adjustment is first performed according to the user real-time data, so that the light source can better meet the actual needs of the user, avoid the light source from affecting the user's skin color, etc., and improve the accuracy of the user's face three-dimensional model construction, so that the final generated simulated makeup effect is more in line with the actual user situation.
[0012] Furthermore, generating a user makeup effect using a user facial three-dimensional model and preset makeup data includes: updating the user facial three-dimensional model using the preset makeup data to obtain a simulated three-dimensional model; and rendering a two-dimensional image using the simulated three-dimensional model to generate a simulated makeup effect.
[0013] Furthermore, it also includes: obtaining user process data, updating the user's facial three-dimensional model according to the process data to obtain a real-time three-dimensional model; and outputting makeup difference data using the real-time three-dimensional model and the simulated three-dimensional model.
[0014] Furthermore, it also includes: building a product database based on user product information; obtaining product usage frequency based on user historical product usage data; and outputting a personalized reminder strategy based on the product database and product usage frequency.
[0015] Another technical solution provided by the present application is a light source adjustment device, which is used to implement the light source adjustment method as described above, and is connected to the light source module, including: a data acquisition module, which is used to obtain the object to be identified; a difference analysis module, which is used to obtain a color difference value based on the difference between the real-time color parameters of the object to be identified and the preset color parameters; and a control and adjustment module, which is used to perform light source module adjustment according to the color difference value.
[0016] Another technical solution provided by the present application is a beauty mirror, which is used to implement the intelligent beauty method as mentioned above, and at least includes a mirror body and a light source module, and also includes: a data acquisition module, which is used to collect user real-time data and user updated data; a data modeling module, which is used to build a user's facial three-dimensional model according to the user's updated data; an interactive analysis module, which pre-stores preset makeup data, and is used to generate simulated makeup effects according to the user's facial three-dimensional model and the preset makeup data, and interact with the user according to the simulated makeup effects.
[0017] Another technical solution provided by the present application is a beauty mirror, which is used to implement the above method, including: a light source module, which is used to provide light source illumination; a product database, which is used to pre-store user product information; and a control analysis module, which is used to output simulated makeup effects according to user real-time data and user beauty data. The difference analysis module, the control adjustment module, and the control analysis module are integrated into a control chip, thereby reducing the space occupied by the control system. There is no need to add an additional control system for regulating the light source module on the beauty mirror. Through the integration of AR facial modeling and light source adjustment, the accuracy of the final output simulated makeup effect is improved, and the impact on the original space of the beauty mirror is avoided.
[0018] Beneficial effects of this application: 1. According to the real-time color parameters and preset color parameters of the object to be identified, the difference between the color of the object to be identified under the current light source and its actual color is obtained, and a color difference value is obtained based on the color difference and the light source adjustment is performed accordingly, until the color difference value is reduced to a preset range, so that the color of the object to be identified under the light source conditions is close to the actual color, without the need to perform spectral identification and comparison of the light source light, reducing the amount of calculation, improving the adjustment efficiency and being able to adapt to the actual reaction of the human eye to light.
[0019] 2. Obtain user real-time data, and first perform light source adjustment based on the user's real-time data to make the light source better meet the user's actual needs, avoid the light source from affecting the user's skin color, etc., improve the accuracy of building the user's facial three-dimensional model, and make the final simulated makeup effect more in line with the actual user situation.
[0020] 3. According to the color of the center point and the foreground color of the transition boundary, the influence of the background color on the foreground color of the transition boundary is obtained, that is, the color compensation value, so as to reduce the influence of the light source on the color compensation value in the construction of the color compensation model and avoid over-compensation.
[0021] 4. First, based on the user's skin type information, such as dry, oily, combination, sensitive, etc., select products suitable for the skin type from the second recommended product information in the product database, and further analyze the ingredients and efficacy of these products to find products that are similar to the products currently used by the user or can solve the user's specific skin problems, such as spots, wrinkles, acne, etc., and present them to the user to achieve personalized recommendations.
[0022] 5. Integrate the difference analysis module, control adjustment module and control analysis module into one control chip, thereby reducing the space occupied by the control system. There is no need to add an additional control system for regulating the light source module on the beauty mirror. Through the integration of AR facial modeling and light source adjustment, the accuracy of the final output simulated makeup effect can be improved, avoiding the impact on the original space of the beauty mirror. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 This is a flow chart of the light source adjustment method of the present application.
[0024] Figure 2 This is a flowchart of the intelligent beauty method of this application. DETAILED DESCRIPTION
[0025] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application is further described in detail below in conjunction with the drawings and examples. It should be understood that the specific implementation method described here is only an optimal embodiment of the present application, which is only used to explain the present application and does not limit the scope of protection of the present application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.
[0026] like Figure 1 As shown, as the first embodiment of the present application, the light source adjustment method includes the following steps: Acquire the object to be identified, confirm the preset color parameters of the object to be identified, and identify the real-time color parameters of the object to be identified; Obtaining a color difference value based on a difference between the real-time color parameter and the preset color parameter; Light source adjustment is performed according to the color difference value until the color difference value is reduced to a preset range.
[0027] In this embodiment, based on the real-time color parameters and preset color parameters of the object to be identified, the difference between the color of the object to be identified under the current light source and its actual color is obtained, and a color difference value is obtained based on the color difference and the light source adjustment is performed accordingly until the color difference value is reduced to a preset range, so that the color of the object to be identified under the light source conditions is close to the actual color, and there is no need to perform spectral identification and comparison of the light source light, which reduces the amount of calculation, improves the adjustment efficiency and can adapt to the actual reaction of the human eye to light.
[0028] The object to be identified can be an object or human body that reflects the light source condition. The original colors of each category of objects are pre-stored as preset color parameters in the database. When the light source adjustment is performed, the object to be identified is obtained, and the database is called to confirm the preset color parameters of the object to be identified. The preset color parameters at least include preset two-dimensional RGB values, and the real-time color parameters at least include real-time two-dimensional RGB values. At this time, the color difference value obtained based on the difference between the real-time color parameters and the preset color parameters includes: The RGB difference value is calculated according to the preset two-dimensional RGB value and the real-time two-dimensional RGB value as the color difference value.
[0029] In this case, the object to be identified is obtained, and the RGB values of different areas on the two-dimensional image of the object to be identified are obtained as real-time two-dimensional RGB values, and the RGB difference is calculated according to the preset two-dimensional RGB value and the real-time two-dimensional RGB value, and then the light source is adjusted according to the RGB difference, so that the color parameters of the object after adjustment are close to the preset color parameters. It can be understood that due to factors such as human eye errors and spatial position differences, it is impossible to achieve that the color parameters of the object after adjustment are completely consistent with the preset color parameters in practice, so a preset range is set, that is, when the color difference value is reduced to the preset range, it can be considered that the current light source adjustment is completed. For example, under the current light source, the color rendering index of red is 90, and the preset range is 10. At this time, the color rendering index of red, that is, the color parameter, has met the requirements, but the color rendering index of yellow is 70, which does not meet the preset range. At this time, it is still necessary to continue to adjust the color temperature of the light source so that the color difference values of different colors meet the preset range, thereby avoiding the influence of the color temperature difference of the light source on the color display.
[0030] In a two-dimensional image, the color of each pixel can be represented by the values of the three RGB channels. After acquiring the object to be identified, a two-dimensional image of the object to be identified is obtained, and the RGB value of each pixel is obtained according to the two-dimensional image of the object to be identified. According to the preset two-dimensional RGB value corresponding to the pixel, the RGB difference of each pixel is calculated as the color difference value, and the RGB difference of each pixel is reduced to a preset range according to the color difference value.
[0031] In other cases, the preset range includes a preset ratio threshold and a preset difference threshold, and the light source adjustment is performed according to the color difference value until the color difference value is reduced to a preset range including: Perform light source adjustment according to the color difference values of all pixels; When the ratio of the color difference value being less than the preset difference threshold reaches the preset ratio threshold, the light source adjustment is completed.
[0032] Due to differences in actual conditions, it is usually impossible to completely match all pixels. At this time, a preset ratio threshold and a preset difference threshold are set. When the difference between the real-time color parameters of a pixel and the preset color parameters is less than the preset difference threshold, it is considered that the color parameters of the pixel under the current light have met the preset color parameters. At this time, the number of pixels that meet the requirements is counted. If the ratio of the number of pixels to the total number of pixels reaches the preset ratio threshold or is greater than the preset ratio threshold, it is considered that the real-time color parameters of the object to be identified and the preset color parameters have reached the required matching degree, and the light source adjustment is completed. It can be understood that during the traversal calculation process, if the ratio of the current color difference value to the preset difference value has exceeded the complementary value of the preset ratio threshold, it is considered that the light source adjustment needs to be continued, and the real-time color parameters of the remaining pixels that have not been traversed are no longer compared, thereby improving the adjustment efficiency.
[0033] Performing light source adjustment according to the color difference value until the color difference value is reduced to a preset range also includes: When the ratio of the color difference value less than the preset difference value threshold does not reach the preset ratio threshold, determine whether the ratio under the previous adjustment is greater than the ratio under the current adjustment. If not, execute the current light source adjustment strategy. If so, execute the current light source adjustment strategy in reverse.
[0034] For example, when adjusting the light source, there will be an initial light source adjustment strategy, such as brightening or dimming. At this time, if the current light source adjustment strategy is to enhance the brightness, and the proportion of the color difference value under the current adjustment is less than the preset difference value threshold is smaller than the proportion of the color difference value under the previous adjustment is less than the preset difference value threshold, then the current adjustment does not meet expectations, and the current light source adjustment strategy is executed in reverse, that is, the brightness is reduced, and the proportion of the color difference value under the current adjustment is less than the preset difference value threshold is greater than the proportion of the color difference value under the previous adjustment is less than the preset difference value threshold, indicating that the light source adjustment is close to the demand situation, then the light source adjustment continues to be performed according to the strategy of enhancing brightness.
[0035] In other embodiments, the preset color parameters at least include preset HSV color space parameters, and the real-time color parameters at least include real-time HSV color space parameters. In this case, the color difference value obtained based on the difference between the real-time color parameters and the preset color parameters includes: The HSV color space difference is calculated according to the preset HSV color space parameter and the real-time HSV color space parameter as the color difference value.
[0036] The HSV color space consists of three components: hue (H), saturation (S), and value (V). Hue indicates the category of color, such as red, green, blue, etc.; saturation indicates the depth of color; and value indicates the brightness of color. Compared with the RGB color space, the HSV color space is more in line with the perception of the human visual system. By distinguishing the real-time parameters and preset parameters of the HSV color space, the color of the object to be identified after the light source is adjusted can be made more in line with the perception of the human visual system, so that the adjustment result is more in line with human perception. It can be understood that in the process of color recognition, the two-dimensional image is converted from the RGB color space to the HSV color space, that is, the actual HSV color space parameters can be converted from the actual two-dimensional RGB values.
[0037] As a second embodiment of the present application, the real-time color parameters also include real-time background color parameters and real-time foreground color parameters. At this time, the color difference value obtained based on the difference between the real-time color parameters and the preset color parameters also includes: Screening real-time color parameters to obtain real-time background color parameters and real-time foreground color parameters; Performing color compensation according to the real-time background color parameter and the real-time foreground color parameter to obtain updated color parameters; The color difference value is obtained by the difference between the updated color parameter and the preset color parameter.
[0038] Due to the color contrast effect, the background color will affect the foreground color. Therefore, compensation is made based on the impact of the background color on the foreground color, and the real-time foreground color parameters are updated to obtain updated color parameters. Then, the impact of the background color is excluded when comparing the preset color parameters, avoiding over-adjustment of the light source and improving the accuracy of light source adjustment.
[0039] In this embodiment, screening the real-time color parameters to obtain the real-time background color parameters and the real-time foreground color parameters includes: The real-time background color parameter and the first real-time foreground color parameter located at the transition boundary in the real-time color parameter are obtained.
[0040] At this time, the light source adjustment method further includes: Based on the historical training picture, a background color parameter located at a transition boundary, a first foreground color parameter, and a second foreground color parameter located at a center point in the historical training picture are obtained; According to the matching of the first foreground color parameter and the second foreground color parameter, a color compensation value corresponding to the background color parameter is obtained, and a color compensation model is constructed with the background color parameter, the first color parameter and the color compensation value.
[0041] The updated color parameters obtained by color compensation based on the real-time background color parameters and the real-time foreground color parameters also include: The real-time background color parameter and the first real-time foreground color parameter are used to output an updated color parameter using a color compensation model.
[0042] Because usually, when taking pictures or videos of the object to be identified, the background image will be captured, and the background image will affect the color condition of the foreground image due to the color contrast effect. Therefore, through the historical training pictures, the color of the center point is matched with the foreground color at the transition boundary. It can be understood that, according to habit, the object to be identified is usually placed in the middle area of the shot, that is, the color of the center point is usually the color of the object to be identified itself, and because the center point is far away from the background color, the influence of the color contrast effect can be ignored. It is believed that the color of the center point is only affected by the light source at this time. According to the matching of the color of the center point with the foreground color of the transition boundary, the foreground color of the transition boundary is affected by the background color, that is, the color compensation value, thereby reducing the influence of the light source on the color compensation value in the construction of the color compensation model to avoid over-compensation.
[0043] As a third embodiment of the present application, a light source adjustment device is connected to the light source module, including: A data acquisition module, used to obtain the object to be identified; The difference analysis module is used to obtain a color difference value according to the difference between the real-time color parameters of the object to be identified and the preset color parameters; the control and adjustment module is used to perform light source module adjustment according to the color difference value.
[0044] In this embodiment, the data acquisition module is connected to the difference analysis module, the difference analysis module is connected to the control and regulation module, and the control and regulation module is connected to the light source module.
[0045] The data acquisition module at least includes a camera for taking pictures of the object to be identified. In other cases, the data acquisition module also includes a scanner.
[0046] The difference analysis module includes at least an image processing module and a difference calculation module. The image processing module receives a picture of an object to be identified, performs type recognition of the object to be identified, thereby confirming the preset color parameters of the object to be identified, and identifies the real-time color parameters of the object to be identified. The difference calculation module receives the preset color parameters and real-time color parameters output by the image processing module, and performs color difference value calculation. In this embodiment, the image processing module includes at least a CMOS sensor for realizing image digitization to facilitate subsequent color recognition. The image processing module pre-stores image feature extraction and classification algorithms to facilitate the recognition of the type of object to be identified.
[0047] A preset range is pre-stored in the control adjustment module. During the adjustment process, the adjustment amplitude and direction of the light source are adjusted according to the changes in the color difference value.
[0048] In other embodiments, the image processing module also includes a transition boundary recognition unit and a compensation unit, which performs background color parameter and first foreground color parameter recognition of the transition boundary according to the transition boundary recognition unit, and outputs corresponding color compensation values through a color compensation model constructed by the compensation unit based on historical training pictures.
[0049] like Figure 2 As shown, as the fourth embodiment of the present application, the intelligent makeup method includes the following steps: Acquire user real-time data and perform light source adjustment based on the user real-time data; Obtaining user update data after light source adjustment, and constructing a user face 3D model according to the user update data; Generate simulated makeup effects based on the user's 3D facial model and preset makeup data.
[0050] In this embodiment, the user's real-time data is obtained, and the light source is first adjusted according to the user's real-time data, so that the light source can better meet the user's actual needs, avoid the light source from affecting the user's skin color and other aspects, improve the accuracy of constructing the user's facial three-dimensional model, and make the final generated simulated makeup effect more in line with the actual user situation.
[0051] In this embodiment, generating a user makeup effect using a user's facial 3D model and preset makeup data includes: Updating the user's facial three-dimensional model with the preset makeup data to obtain a simulated three-dimensional model; Rendering a two-dimensional image with a simulated three-dimensional model generates a simulated makeup effect.
[0052] The user's facial 3D model is updated according to the preset makeup data, and the makeup is overlaid in 3D form to be close to the user's actual facial makeup effect. This is then used for 2D image rendering to generate a simulated makeup effect, which is output in the form of a 2D image that is more in line with the 3D situation. This not only makes it easier for users to view the makeup effect, but also shows the user's photo-taking effect with the current makeup, thereby improving the adaptability to user needs.
[0053] The smart beauty approach also includes: Acquire user process data, and update the user's facial 3D model according to the process data to obtain a real-time 3D model; Output makeup difference data using real-time 3D models and simulated 3D models.
[0054] When the user is applying makeup according to the simulated makeup effect, real-time feedback is required as to whether there is a difference between the real-time makeup effect and the simulated makeup effect, thereby obtaining the user's process data, updating the user's facial 3D model according to the process data to obtain a real-time 3D model, and performing makeup comparison on the 3D model to avoid errors caused by rendering.
[0055] In this embodiment, the intelligent makeup method further includes: When the makeup difference data exceeds the preset makeup difference threshold, a makeup deviation prompt is given.
[0056] The user is prompted only when the makeup difference data exceeds the preset makeup difference threshold, avoiding frequent prompts for minor errors that do not affect the actual makeup effect, thereby improving the user experience.
[0057] In this embodiment, the user real-time data includes at least the user's real-time facial data, and the user process data includes at least the user's real-time makeup data. Professional makeup steps and techniques are presented to the user in the form of three-dimensional images through AR technology, and combined with voice prompts or text instructions, to help users learn and master makeup techniques.
[0058] As a fifth embodiment of the present application, obtaining the user updated data after the light source is adjusted, and constructing the user face three-dimensional model according to the user updated data further includes: Extract features and classify and mark the user update data to obtain several marked feature points; Construct a 3D facial model of the user based on several marked feature points.
[0059] In this embodiment, the user update data includes at least user image data or video data after light source adjustment. Feature extraction and classification marking are performed on the user update data to obtain a number of marked feature points. A user facial 3D model is constructed based on the number of marked feature points. In some cases, the user facial 3D model can be stored as a configuration model. The user facial 3D models of multiple users are pre-stored. When using AR-assisted makeup, the corresponding user facial 3D model is called, which is suitable for users to independently select simulated makeup effect output objects and improve user experience.
[0060] In this embodiment, updating the user's facial 3D model with the preset makeup data to obtain a simulated 3D model further includes: Perform regional analysis based on preset makeup data to obtain facial regions; Perform key point analysis on the facial features to obtain the facial features coverage area; Perform key point analysis on the makeup part of the facial area to obtain the makeup coverage area; Perform a proportional analysis of the makeup coverage area based on the facial features coverage area to obtain coverage parameters; The color features of the makeup coverage area are extracted, and the user's facial 3D model is updated with the color features, coverage parameters and makeup coverage area to obtain a simulated 3D model.
[0061] By analyzing the proportions of facial features and makeup coverage, we can obtain the coverage parameters of makeup on the facial features, and then use color features, coverage parameters and coverage areas to cover the user's facial three-dimensional model to obtain a simulated three-dimensional model, thereby improving the accuracy of makeup simulation.
[0062] The override parameters are: Among them, FCS i is the i-th makeup coverage parameter of the preset makeup data, SZR i is the area covered by the i-th makeup in the preset makeup data, SWG i is the area covered by the i-th facial features of the preset makeup data.
[0063] Updating the user's facial 3D model with color features, coverage parameters, and makeup coverage areas to obtain a simulated 3D model also includes: Constructing a first makeup model according to the coverage parameters and the makeup coverage area; Rendering the first makeup model according to the color feature to obtain a second makeup model; The second makeup model is fused to the corresponding position of the user's facial three-dimensional model to obtain a simulated three-dimensional model.
[0064] As the sixth embodiment of the present application, the intelligent makeup method further includes: Build a product database based on user product information; Obtain product usage frequency based on user historical product usage data; Output personalized reminder strategies based on product database and product usage frequency.
[0065] User product information includes at least the category of beauty products and the validity period of beauty products. User historical product usage data includes at least the historical usage times corresponding to the historical beauty product usage category. Users need to use up all beauty products before the expiration date to avoid waste. Therefore, it is possible to calculate whether the current product usage frequency can be used within the validity period based on the expiration date and the current product usage frequency. If not, the corresponding beauty products are added to the personalized reminder strategy to prompt users to use them as soon as possible to avoid waste.
[0066] In this embodiment, the intelligent makeup method further includes: Obtain user historical data and use collaborative filtering algorithms to obtain similar user behavior data that matches the user historical data; Output recommended product information to the product database based on similar user behavior data.
[0067] User history data at least includes user purchase history, user browsing behavior, and user evaluation. Collaborative filtering is a recommendation algorithm based on user behavior. In the scenario of cosmetics recommendation, by analyzing the user's purchase history, browsing behavior, evaluation and other data, other users with similar interests to the user are found, and then based on the behavior of these similar users, product data that better meets the user's needs, namely recommended product information, is obtained and stored in the product database.
[0068] In this case, when the makeup difference data exceeds the preset makeup difference threshold, the makeup deviation prompt also includes: when the makeup difference data exceeds the preset makeup difference threshold, a makeup position deviation prompt and a makeup color deviation prompt are provided; and the corresponding recommended product information is output according to the makeup color deviation prompt and the product database.
[0069] Since the user's existing beauty products may have color differences from the preset makeup data, resulting in makeup deviation, when there is a makeup color deviation, a makeup color deviation prompt will be given, and the recommended product information pre-stored in the product database will be matched to show more suitable beauty products to the user.
[0070] In other embodiments, the intelligent makeup method further includes: Output the corresponding recommended product information based on the user's facial 3D model and product database.
[0071] The user's skin quality information and skin problems are obtained based on the user's facial 3D model, and the recommended product information in the product database is matched with the skin quality information and skin problems to obtain recommended product information that is suitable for the skin type and can solve skin problems, and personalized recommendations are made to the user.
[0072] In this embodiment, the recommended product information includes at least information such as product ingredients, product efficacy, and applicable skin types.
[0073] As the seventh embodiment of the present application, the intelligent makeup method further includes: Obtain user historical data and use collaborative filtering algorithms to obtain similar user behavior data that matches the user historical data; Outputting first recommended product information to a product database according to similar user behavior data; Obtain product data and use content filtering algorithms to obtain product features; Output the second recommended product information to the product database based on the user's historical data and product characteristics.
[0074] The first recommended product information is the recommended product in similar user behavior data, and the second recommended product is the recommended product output based on the characteristics of the product itself to meet the user's historical usage habits. Content filtering is based on the characteristics of the cosmetic product itself and the user's personal preferences. A detailed analysis of the attributes of cosmetic ingredients, efficacy, applicable skin type, etc. is performed to determine whether the characteristics of the product itself fit the user's preferences. It can be understood that product data can be obtained through big data, that is, all existing product data can be obtained through the cloud database, and content filtering is performed to obtain the corresponding product characteristics. First, based on the user's skin type information, such as dry, oily, mixed, sensitive, etc., select products suitable for the skin type from the second recommended product information in the product database, and further analyze the ingredients and efficacy of these products to find products that are similar to the products currently used by the user or can solve the user's specific skin problems, such as spots, wrinkles, acne, etc., and present them to the user to achieve personalized recommendations.
[0075] It is understandable that in actual applications, collaborative filtering and content filtering are usually used in combination to fully utilize the advantages of both methods and improve the accuracy of recommendations.
[0076] The recommended product information outputted based on the user's facial 3D model and the product database also includes: Based on the positioning and weather information, the corresponding recommended product information is output according to the user's facial 3D model and the product database.
[0077] In this case, considering that different geographical locations and seasons have different applicable products, the corresponding recommended product information is output based on the location and weather information. At this time, the product data at least includes ingredients, efficacy, applicable skin type, applicable age, applicable gender, and applicable geographical location. It is understandable that the user's age and gender can be directly obtained based on the user's facial 3D model.
[0078] Outputting the second recommended product information to the product database according to the user history data and the product characteristics also includes: The similarity between the user's historically used products and the product to be recommended is calculated based on the product characteristics, and the second recommended product information is output to the product database based on the similarity.
[0079] Calculate the similarity between cosmetic products, compare and match the cosmetic attributes, and output the second recommended product information for products with similar attributes to the products the user has used in the past but with more suitable effects. It can be understood that the products to be recommended are the existing cosmetic products obtained from the cloud database.
[0080] Similarly, based on the user's skin type and the information in the product database, it is possible to determine whether the product currently used by the user has problems with being incompatible with his or her skin type or prone to causing adverse reactions. If so, the user will be prompted.
[0081] As an eighth embodiment of the present application, the intelligent makeup method further includes: Monitor the dust conditions in the space where beauty products are located in real time, and issue cleaning reminders when the dust conditions reach the set dust threshold.
[0082] In this embodiment, by real-time monitoring of the dust situation in the space where the cosmetic products are located, the user is reminded to clean it, so as to avoid excessive dust in the space affecting the user's experience, and at the same time avoid the dust of a certain cosmetic product affecting other cosmetic products. It can be understood that the cosmetic products here at least include cosmetic tools, such as makeup brushes, eye shadow brushes, etc., which will carry makeup powder after use. If they are not cleaned in time, the cosmetic tools will be contaminated with different makeup powders, thereby affecting the makeup experience of the end user.
[0083] As the ninth embodiment of the present application, a beauty mirror includes: Data collection module, used to collect user real-time data and user update data; A data modeling module, used to construct a user's facial 3D model based on the user's updated data; The interactive analysis module pre-stores preset makeup data, is used to generate simulated makeup effects based on the user's facial 3D model and the preset makeup data, and interact with the user based on the simulated makeup effects.
[0084] The data modeling module is connected to the interactive analysis module, and outputs simulated makeup effects for users to browse by constructing a user facial model and overlaying preset makeup data.
[0085] The preset makeup data can cover various styles from daily makeup to professional makeup. Users can browse and select their favorite makeup through the interactive analysis module of the beauty mirror.
[0086] Specifically, the beauty mirror at least includes a mirror body and a light source module for performing light source adjustment. The interactive analysis module at least includes a touch screen arranged on the mirror body, and a touch screen integrated on the mirror body, which is used to display skin analysis results, beauty tutorials, virtual makeup trial effects, etc., and provides an interface for users to interact with the beauty mirror. The light source module at least includes an LED light strip arranged around the mirror body. The data acquisition module at least includes a camera arranged on the mirror body. The user's photos or videos are taken by the camera to obtain the user's real-time data. The light source module adjusts the light source based on the user's real-time data, obtains the user's updated data after adjustment, and constructs a three-dimensional model of the user's face through the data modeling module. The interactive analysis module generates simulated makeup effects and displays them on the touch screen. Users can select different simulated makeup effects through the touch screen.
[0087] In this embodiment, the beauty mirror also includes a data communication module, which is connected to the cloud database and the interactive analysis module. The cloud database is the product database, which is connected to the product database to identify and import product information, including product ingredients, purchase date, shelf life, etc., and manage according to system management suggestions and usage period, and output product information based on the cloud database to maximize the beauty and skin care effect.
[0088] At the same time, the data communication module is also connected to the product database and the data acquisition module, and users can upload product information to the product database through the data acquisition module.
[0089] In this embodiment, the data communication module at least includes: a wifi / bluetooth communication module.
[0090] The beauty mirror also includes a storage space for storing beauty products, and the data acquisition module also includes a dust sensor. The dust situation in the storage space is collected by the dust sensor, and when the dust situation in the storage space reaches the set dust threshold, a cleaning reminder is issued. The storage space is arranged at the rear end or below the mirror body, and the dust sensor is arranged in the storage space. The detection principle of the dust sensor mainly relies on optical or electrical methods. Specifically, the optical method usually adopts the principle of laser scattering, and there is a laser light source and a photosensitive device inside the sensor. When dust particles enter the detection area of the sensor, they scatter the laser beam, and the scattered light is then received by the photosensitive device. By measuring the intensity of the received scattered light, the sensor can calculate the dust concentration in the air.
[0091] As the tenth embodiment of the present application, obtaining user real-time data, and performing light source adjustment based on the user real-time data also includes: obtaining user real-time data, using the user real-time data as the object data to be identified, and performing light source adjustment using a light source adjustment method.
[0092] In this embodiment, since the real-time data of the user is also obtained by shooting with a camera, the object to be identified can be obtained by using the user image captured by the camera. At this time, the object to be identified is the user. The preset color parameters are confirmed according to the skin color of the user, and the skin color of the user under the current light source is identified as the real-time color parameter. Then, the color difference is calculated according to the actual skin color of the user and the skin color under the current light source, and the light source is adjusted so that the skin color of the user under the current light source is closer to the actual skin color of the user. The actual skin color of the user can be obtained according to the skin color under natural light, or it can be set by the user. It is understandable that although the skin color of the human body does not change due to environmental changes, since the color difference is based on image processing technology and human perception is also affected by the environment, the skin color will show different perceived colors under different light sources, such as the skin color will be darker when the light source is darker, and the skin color will be brighter when the light source is brighter. The brightness and darkness of the facial features of a three-dimensional face under different light sources will also affect the presentation of the makeup color. Therefore, the light source is adjusted according to the skin color of the human body, so that the light source is closer to the actual situation, which is convenient for the accurate presentation of the makeup. At the same time, based on the color difference value calculation obtained by image processing, there is no need for additional hardware design, which reduces the hardware loading on the beauty mirror, reduces the manufacturing requirements of the beauty mirror, meets the light source adjustment requirements, reduces hardware costs, improves the portability of the beauty mirror, and thus improves the practicality of the beauty mirror and facilitates its application.
[0093] In other embodiments, the object to be identified may also be a product used by the user, that is, to identify the beauty products of the user. Usually, beauty products have fixed color numbers, and preset color parameters can be constructed through the color characteristics of different color numbers under natural light, and then the light source is adjusted to make the color displayed by the beauty products more consistent with the color under natural light conditions, so that the user can more intuitively feel the display effect of the current makeup under natural light during the makeup process. Of course, in this embodiment, considering that most users have higher requirements for makeup adaptation under natural light, and in special cases, such as as a stage beauty mirror, the makeup effect it displays should be more in line with stage lighting, and preset color parameters under multiple light sources can be pre-stored to provide users with more choices to adapt to the different usage needs of customers.
[0094] In order to further distinguish the objects to be identified, a recognition area can be preset. When the user places the beauty product in the preset recognition area, the beauty product recognition is performed. When the user does not place the beauty product in the preset recognition area, the user's skin color recognition is performed. This adapts to the different needs of users.
[0095] As an eleventh embodiment of the present application, a beauty mirror includes: A data acquisition module, used to obtain the object to be identified; A difference analysis module is used to obtain a color difference value according to the difference between the real-time color parameter of the object to be identified and the preset color parameter; a control and adjustment module is used to perform light source module adjustment according to the color difference value; A light source module, used for providing light source illumination; Product database, used to pre-store user product information; The control and analysis module is used to output simulated makeup effects based on the user's real-time data and user's beauty data.
[0096] In the present embodiment, the data acquisition module obtains the object to be identified, that is, obtains the real-time data of the user, and performs light source adjustment through the data acquisition module that originally performs the output of simulated makeup effects of the beauty mirror and the product database. It can be understood that although the difference analysis module and the control and adjustment module are separated separately in the present embodiment, based on the integration capability of the control chip, it is possible to integrate the difference analysis module, the control and adjustment module, and the control analysis module into one control chip, thereby reducing the space occupied by the control system, and there is no need to add an additional control system for regulating the light source module on the beauty mirror. Through the integration of AR facial modeling and light source adjustment, the impact on the original space of the beauty mirror can be avoided while improving the accuracy of the final output simulated makeup effect.
[0097] Specifically, the product database includes at least user product information and recommended product information. The user product information includes at least user product specifications, user product color numbers and user product images. According to the user's rejected product images and the image information taken in the preset recognition area, it is possible to match which beauty product the object to be recognized is, and then call the preset color parameters according to the color number of the corresponding beauty product.
[0098] In this embodiment, the beauty mirror also includes a data communication module, which is connected to the difference analysis module, the control analysis module and the product database. The data communication module is also connected to the product database and the cloud network. It can obtain existing beauty product information from the cloud network and obtain recommended product information based on collaborative filtering and content filtering.
[0099] The data communication module is also connected to the data acquisition module. When the mirror is connected to the Internet, the barcode label of the product is scanned by the camera, and the cloud network is automatically called to obtain detailed product information. The cloud network can be an interface of the product information service platform. In other embodiments, the cloud network can also be a meteorological website or other Internet service website. The data communication module communicates with smart devices such as smart phones and Internet services to achieve data upload and download and remote control. The weather environment data is obtained in real time through the information acquisition function of the Internet, and then the corresponding preset color parameters are automatically selected according to the current external weather conditions to correct and adjust the mirror light source, so as to present the user with a makeup experience that is closest to the actual use effect. For example, the weather, ultraviolet rays, Guangchao and other information of the user's location are automatically obtained through the Internet every hour to adapt to the user's outdoor makeup needs.
[0100] In this embodiment, the beauty mirror also includes a storage space for storing beauty products, and the data acquisition module also includes a dust sensor. Through the establishment of a product database, personalized suggestions can be provided for the use of beauty products and makeup of users, and the light source can be automatically corrected by using product characteristics, the accuracy of makeup simulation display can be improved, and color difference can be reduced, so that users can conduct virtual makeup trials, which helps to meet the basic skin care and beauty needs of users and help users to more scientifically skin care and beauty, thereby improving the user experience. It can effectively reduce the user's repeat purchase rate, improve the cleanliness of the user's beauty products, further enhance the user's beauty experience, meet the user's needs for makeup simulation in different situations, and can not only simulate the current user's own makeup effect, but also simulate the makeup effect of the user of the pre-stored facial three-dimensional model, which is more suitable for multi-user use.
[0101] The specific implementation methods described above are preferred implementation methods of the light source adjustment method and device, intelligent beauty method and beauty mirror of the present application, and are not intended to limit the specific implementation scope of the present application. The scope of the present application includes but is not limited to the specific implementation methods. All equivalent changes made in accordance with the shape and structure of the present application are within the protection scope of the present application.
Claims
1. A light source adjustment method, characterized in that: The steps include: Acquire the object to be identified, confirm the preset color parameters of the object to be identified, and identify the real-time color parameters of the object to be identified; Obtaining a color difference value based on a difference between the real-time color parameter and the preset color parameter; Light source adjustment is performed according to the color difference value until the color difference value is reduced to a preset range.
2. The light source adjustment method according to claim 1, characterized in that: The step of obtaining a color difference value based on the difference between the real-time color parameter and the preset color parameter further includes: Screening real-time color parameters to obtain real-time background color parameters and real-time foreground color parameters; Performing color compensation according to the real-time background color parameter and the real-time foreground color parameter to obtain updated color parameters; The color difference value is obtained by the difference between the updated color parameter and the preset color parameter.
3. The light source adjustment method according to claim 1, wherein: Also includes: Based on the historical training picture, a background color parameter located at a transition boundary and a first foreground color parameter and a second foreground color parameter located at a center point in the historical training picture are obtained; According to the matching of the first foreground color parameter and the second foreground color parameter, a color compensation value corresponding to the background color parameter is obtained, and a color compensation model is constructed with the background color parameter, the first color parameter and the color compensation value.
4. An intelligent makeup method, characterized in that: The steps include: Acquire user real-time data, and execute the light source adjustment method according to any one of claims 1 to 3 based on the user real-time data; Obtaining user update data after light source adjustment, and constructing a user face 3D model according to the user update data; Generate simulated makeup effects based on the user's 3D facial model and preset makeup data.
5. The intelligent makeup method according to claim 4, characterized in that: The generating of the user makeup effect by using the user's facial 3D model and the preset makeup data comprises: Updating the user's facial three-dimensional model with preset makeup data to obtain a simulated three-dimensional model; Rendering a two-dimensional image with a simulated three-dimensional model generates a simulated makeup effect.
6. The intelligent makeup method according to claim 5, characterized in that: Also includes: Obtain user process data, and update the user's facial 3D model according to the process data to obtain a real-time 3D model; Output makeup difference data using real-time 3D models and simulated 3D models.
7. The intelligent makeup method according to claim 4, characterized in that: Also includes: Build a product database based on user product information; Obtain product usage frequency based on user historical product usage data; Output personalized reminder strategies based on product database and product usage frequency.
8. A light source adjustment device, used to implement the light source adjustment method according to any one of claims 1 to 3, connected to a light source module, characterized in that: include: A data acquisition module, used to obtain the object to be identified; A difference analysis module, used to obtain a color difference value according to the difference between the real-time color parameter of the object to be identified and the preset color parameter; The control and adjustment module is used to perform light source module adjustment according to the color difference value.
9. A makeup mirror, used to implement the intelligent makeup method according to any one of claims 4 to 7, comprising at least a mirror body and a light source module, characterized in that: Also includes: Data collection module, used to collect user real-time data and user update data; A data modeling module, used to construct a user's facial 3D model based on the user's updated data; The interactive analysis module pre-stores preset makeup data, is used to generate simulated makeup effects based on the user's facial 3D model and the preset makeup data, and interact with the user based on the simulated makeup effects.
10. A makeup mirror, used to implement the method according to any one of claims 8, characterized in that: include: A light source module, used for providing light source illumination; Product database, used to pre-store user product information; The control analysis module is used to output simulated makeup effects based on the user's real-time data and user beauty data.
