A multifunctional cosmetic compact

Abstract

The application discloses a multifunctional cosmetic powder box, comprising a skin detection module for collecting skin state data such as skin tone, oil, moisture and texture; an environment perception module for collecting environmental parameter data such as temperature, humidity, illumination and ultraviolet rays; a powder management module for managing the storage and blending of various foundation raw materials; an AI analysis and processing module for generating personalized foundation formula recommendation information based on the skin state data and environmental parameter data using a preset skin-environment-powder mapping model; and an adaptive control module for controlling the powder management module to perform powder blending operations according to the formula recommendation information. The application also integrates wireless communication, user interaction, data storage and power management modules. The application organically integrates skin detection, environment perception, AI analysis and automatic blending functions, realizes a full-process intelligent closed loop from perception to execution, and solves the technical problems of single function, lack of adaptability and low degree of personalization of existing cosmetic powder boxes.

Description

Technical Field

[0001] This invention relates to the field of intelligent beauty equipment technology, and in particular to a multifunctional makeup powder box. Specifically, it relates to a multifunctional makeup powder box system that integrates skin condition detection, environmental factor sensing, personalized foundation formula generation and automatic mixing functions. Background Technology

[0002] Makeup compacts are a widely used beauty tool in daily life, used to hold facial enhancement products such as foundation and powder. With the development of the social economy and the improvement of people's living standards, consumers' demands for cosmetics have upgraded from simply "having" them to "precision" and "personalization," that is, they hope that cosmetics can accurately match their own skin condition, skin tone characteristics, and usage scenarios.

[0003] However, existing cosmetic powder compacts generally suffer from limited functionality. Traditional powder compacts typically serve only as physical containers for foundation or pressed powder, with their "multi-functionality" primarily reflected in structural improvements such as layered designs, adjustable mirror angles, and auxiliary functions like fill lights. For example, Chinese patent CN205963273U discloses a multi-functional cosmetic powder compact that uses a lifting structure to support pressed powder, preventing it from being crushed during replacement. Another example is Chinese patent CN219270349U, which discloses a powder compact with a liftable drawer and a liftable mirror, allowing users to adjust the angle and categorize the powder as needed. While these technological improvements enhance user convenience to some extent, they are essentially limited to physical structural optimizations and do not involve the expansion of intelligent functions.

[0004] Regarding intelligent features, some technical solutions have attempted to integrate electronic modules into cosmetic boxes. For example, Chinese patent application CN201710053974 discloses an intelligent cosmetic box and system, including a cosmetic container and an intelligent base. The intelligent base is equipped with a weight sensor, a main control board, and a wireless communication module. The weight sensor detects changes in the container's weight, which are analyzed, recorded, and displayed to the user by the main control board, allowing the user to control the amount used. This solution achieves accurate measurement of cosmetic usage, but its functionality is still relatively limited, lacking features such as skin condition detection and personalized formula recommendations.

[0005] Chinese patent application CN106021704 discloses a makeup box, including a facial feature recognition module, a mirror display screen, and a wireless communication module. The processor retrieves matching makeup and applicable occasion information from a cloud platform based on facial feature information recognized by the facial feature recognition module and displays it on the mirror display screen. This solution achieves a makeup recommendation function based on facial features, but lacks consideration for environmental factors and does not have the ability to actively mix powder.

[0006] In the field of personalized beauty devices, US patent application US20240177213A1 discloses an AI-based customized cosmetics recommendation device, including a user skin information receiver, a skin determination unit, a skin-customized formula generator, and a customized cosmetics recommendation unit. It determines the user's skin type and condition by analyzing skin images and generates customized formulas. However, this solution primarily focuses on skin analysis and formula recommendation at the software algorithm level and does not deeply integrate with portable makeup powder compact hardware.

[0007] In addition, L'Oréal's Perso concept product uses AI technology to assess consumers' skin condition and combines it with local environmental information to create on-site formulations of skincare and makeup products. The BeautIT AIoT skincare dispenser combines biotechnology, IoT, and AI to achieve real-time skin analysis and precise ingredient dispensing. However, these products are relatively bulky, primarily designed for home use, and lack portability, failing to meet users' needs for touch-ups and personalized makeup anytime, anywhere.

[0008] In summary, the existing technology has the following technical problems that urgently need to be solved: First, the functionality is limited and integration is insufficient. Existing cosmetic powder compacts mainly focus on physical storage and simple functions such as illumination and weighing, lacking systematic integration of multiple functions such as skin condition detection, environmental sensing, and personalized formula generation. There is a lack of organic connection and collaborative working ability between the various functions.

[0009] Secondly, there is a lack of real-time skin condition monitoring capabilities. Users typically rely on experience and visual judgment to determine their skin condition and choose cosmetics during the makeup application process. The lack of objective and quantifiable methods for monitoring skin condition can easily lead to misjudgments, resulting in inappropriate cosmetic selection, affecting makeup application, or even damaging the skin.

[0010] Third, there is a lack of adaptive consideration to environmental factors. Users' skin conditions and makeup needs vary significantly under different environmental conditions (such as temperature, humidity, UV intensity, air quality, etc.). For example, a more moisturizing foundation formula is needed in dry environments, while a higher SPF is required in environments with strong UV radiation. Existing products cannot dynamically adjust cosmetic usage recommendations based on environmental factors.

[0011] Fourth, there is a lack of closed-loop learning and continuous optimization capabilities. A user's skin condition changes dynamically with time, season, and lifestyle habits, and different users have individual differences in their preferences for cosmetics. Existing products lack a self-learning mechanism based on user feedback and usage data, and therefore cannot continuously optimize recommendation accuracy over time, resulting in limited personalization and accuracy of the recommendations.

[0012] Fifth, there is a lack of collaborative control mechanisms for multi-functional modules. Even if some existing products attempt to integrate some intelligent functions, they lack a unified management and scheduling mechanism, resulting in each functional module operating independently, serious information silos, and limitations on the overall system efficiency and user experience.

[0013] Sixth, there is a lack of balance between portability and intelligence. Existing smart beauty devices often increase their size by integrating multiple functions, sacrificing portability; while portable powder compacts lack smart features, forcing users to make trade-offs between "intelligence" and "portability".

[0014] Therefore, there is an urgent need to provide a multifunctional makeup powder box system that can comprehensively solve the above-mentioned technical problems. While maintaining portability, it should organically integrate functions such as real-time detection of skin condition, adaptation to environmental factors, automatic generation and mixing of personalized foundation formulas, and closed-loop learning optimization of usage data, so as to provide users with a one-stop intelligent beauty solution. Summary of the Invention

[0015] To address the shortcomings of existing technologies, the purpose of this invention is to provide a multifunctional cosmetic powder compact that solves the technical problems of existing cosmetic powder compacts, such as limited functionality, lack of skin condition detection capabilities, inability to adaptively adjust to environmental factors, lack of closed-loop learning mechanisms, and the inability to balance portability and intelligence.

[0016] Another objective of this invention is to provide a multifunctional makeup powder box system capable of real-time skin condition detection, environmental factor perception, personalized foundation formula recommendation, and automatic mixing. Through multi-sensor fusion technology and artificial intelligence algorithms, it provides users with scientific, accurate, and personalized makeup solutions.

[0017] Another objective of this invention is to provide a multifunctional cosmetic powder box system with closed-loop self-learning capability, which can continuously optimize the formula recommendation strategy based on the user's actual usage feedback and skin condition change trends, so that the recommendation results gradually approach the user's optimal personalized needs.

[0018] The above-mentioned objective of this invention is achieved through the following technical solutions: This invention provides a multifunctional cosmetic powder box, comprising: A skin detection module, installed on the powder compact, is used to collect the user's skin condition data. The skin detection module includes at least one optical sensor and at least one moisture sensor. The skin condition data includes skin tone data, skin oil content data, skin moisture content data, and skin texture data. An environmental sensing module is installed on the powder compact itself to collect environmental parameter data of the usage environment, including ambient temperature, ambient humidity, ambient light intensity, and ultraviolet index. The powder management module is located inside the cosmetic powder box and is used to manage the storage information and usage records of various foundation raw materials, including at least three different shades and functions of base powder. The AI ​​analysis and processing module is electrically connected to the skin detection module, the environmental perception module, and the powder management module, respectively. It is used to receive the skin state data and the environmental parameter data, perform analysis and processing based on the preset skin-environment-powder mapping model, and generate personalized foundation formula recommendation information. An adaptive control module is also included, which is electrically connected to the AI ​​analysis and processing module and the powder management module, respectively. This module receives the foundation formula recommendation information and controls the powder management module to perform corresponding powder mixing operations based on the foundation formula recommendation information.

[0019] According to one embodiment of the present invention, the skin detection module includes a multispectral image sensor, which includes a visible light imaging unit and a near-infrared imaging unit; the visible light imaging unit is used to acquire visible light images of the user's facial skin to extract skin tone data and skin texture data, and the near-infrared imaging unit is used to acquire deep images of the user's facial skin to extract skin moisture distribution data and subcutaneous feature data.

[0020] Furthermore, the multispectral image sensor also includes an ultraviolet fluorescence imaging unit, which is used to excite specific fluorescent substances on the skin surface to detect the distribution of bacteria and cosmetic residues on the skin surface, providing a basis for skin cleanliness assessment.

[0021] In a preferred embodiment of the present invention, the skin detection module further includes a skin contact sensor group disposed on the outer surface of the cosmetic powder compact body, comprising a capacitive moisture sensor and a photoelectric oil sensor. During use, the user can quickly obtain accurate data on skin moisture and oil content by briefly contacting the facial skin with the skin contact sensor group.

[0022] According to one embodiment of the present invention, the environmental sensing module further includes an ultraviolet sensor and an air quality sensor, and the environmental parameter data further includes ultraviolet intensity and PM2.5 concentration values; the AI ​​analysis and processing module is further used to adjust the sun protection factor parameters and protective efficacy parameters in the foundation formula recommendation information according to the ultraviolet intensity and the PM2.5 concentration values.

[0023] Specifically, when the UV intensity exceeds a preset threshold, the AI ​​analysis and processing module automatically increases the proportion of sunscreen powder in the formula; when the PM2.5 concentration exceeds a preset threshold, the AI ​​analysis and processing module automatically increases the proportion of anti-pollution and antioxidant powder in the formula.

[0024] According to one embodiment of the present invention, the AI ​​analysis and processing module includes: The data preprocessing unit is used to normalize and extract features from the skin state data and environmental parameter data to generate multidimensional feature vectors. The normalization process uses the minimum-maximum normalization method to uniformly map the data of each dimension to the [0, 1] interval to eliminate the bias caused by different units. The feature extraction process includes extracting texture features from the skin image data using the Local Binary Pattern (LBP) algorithm and extracting hue features using the color moment method. The skin condition assessment unit, connected to the data preprocessing unit, is used to determine the user's current skin condition level based on the multidimensional feature vector and a preset skin condition classification model. The skin condition level includes basic types such as dry skin, normal skin, oily skin, and combination skin, as well as subdivided dimensions such as sensitivity and pigmentation level. The formula generation unit is connected to the skin condition assessment unit and the powder management module, respectively. It is used to generate foundation formula recommendation information based on the current skin condition level and the environmental parameter data, combined with the available powder information provided by the powder management module. The foundation formula recommendation information includes the type identifier, ratio parameters, suggested mixing order and usage instructions for each basic powder. The system also includes a cloud collaboration unit, which interacts with the cloud server via a wireless communication module to receive updated skin condition classification models and formula mapping parameters from the cloud. This cloud collaboration unit also uploads locally collected, anonymized usage data to the cloud server, providing data support for the global optimization of the cloud model.

[0025] According to one embodiment of the present invention, the skin condition assessment unit uses a skin condition classification model based on a deep neural network. The input layer of the deep neural network receives the multidimensional feature vector, undergoes nonlinear transformation through multiple hidden layers, and the output layer outputs the confidence distribution of the skin condition level. The structure of the deep neural network includes: an input layer (the dimension of which is the number of dimensions of the multidimensional feature vector), a first fully connected layer (256 neurons, ReLU activation function), a second fully connected layer (128 neurons, ReLU activation function), a Dropout layer (dropout probability of 0.3), a third fully connected layer (64 neurons, ReLU activation function), and an output layer (the number of neurons is the number of skin condition levels, Softmax activation function).

[0026] The multidimensional feature vectors include, but are not limited to: skin tone features (mean and variance of L, a, and b channels based on the L*a*b* color space), skin oil features (oil content values ​​and their gradient changes), skin moisture features (moisture content values ​​and their spatial distribution), skin texture features (texture roughness, orientation, and contrast), ambient temperature features, ambient humidity features, ambient light features, and ultraviolet intensity features.

[0027] In a preferred embodiment of the present invention, the AI ​​analysis and processing module further includes a learning optimization unit. The learning optimization unit is used to: collect user satisfaction feedback data for each generated foundation formula, the satisfaction feedback data being obtained through a user interaction module, including user ratings for dimensions such as color matching, concealing effect, makeup duration, and skin comfort; based on the satisfaction feedback data, combined with corresponding batch skin condition data, environmental parameter data, and actual foundation formula data used, construct a training sample set; and use an incremental learning algorithm to update the parameters of the skin-environment-powder mapping model online, so that the subsequently generated foundation formula recommendation information gradually approaches the user's personalized preferences.

[0028] The incremental learning algorithm preferably employs the Elastic Weight Consolidation (EWC) method. By applying regularization constraints to the key parameters in the model, it avoids catastrophic forgetting of existing knowledge while learning new data, thereby achieving continuous optimization while ensuring model stability.

[0029] According to one embodiment of the present invention, the powder management module includes: At least three powder storage compartments are provided, each for storing base powders of different color numbers or different functions. Each powder storage compartment is equipped with a discharge port and an electrically controlled discharge valve. The electrically controlled discharge valve adopts a screw conveying structure driven by a micro stepper motor or a micro-vibration feeding structure driven by piezoelectric ceramics to achieve precise powder output control. A powder balance monitoring unit is installed in each of the powder storage bins to monitor the powder balance in each powder storage bin in real time. The powder balance monitoring unit adopts a capacitive level sensor or a photoelectric level sensor, which can detect the height or volume change of powder in the powder storage bin in real time, with a monitoring accuracy of ±0.1g. The powder mixing chamber is connected to the outlet of each of the powder storage chambers and is used to receive and mix the powder output from each of the powder storage chambers; the powder mixing chamber is equipped with a micro vibration motor to generate high-frequency micro vibration to promote the uniform mixing of different powders. The system also includes a mixing control unit, which is electrically connected to the electrically controlled discharge valve and the powder balance monitoring unit. This unit receives the allocation instructions from the adaptive control module and controls the opening duration and opening ratio of each of the electrically controlled discharge valves.

[0030] In a preferred embodiment of the present invention, the base powder stored in the powder storage hopper includes: at least one light-colored base powder, at least one intermediate-colored base powder, at least one dark-colored base powder, and at least one functional base powder. The functional base powder includes one or more of the following: a functional powder with sun protection function, a functional powder with oil control function, a functional powder with moisturizing function, a functional powder with antioxidant function, and a functional powder with concealing and enhancing function.

[0031] According to one embodiment of the present invention, the adaptive control module is further configured to: receive the foundation formula recommendation information, parse the type identifier and proportioning parameters of each basic powder contained in the foundation formula recommendation information; generate pulse width modulation control signals for the electrically controlled discharge valves of each powder storage bin in the powder management module according to the type identifier and proportioning parameters; after performing the mixing operation, update the powder usage records in the powder management module, and feed back the usage data to the AI ​​analysis and processing module for optimizing subsequent formula recommendations.

[0032] The duty cycle of the PWM control signal is directly proportional to the target output powder quantity. The higher the duty cycle, the greater the proportion of time the electrically controlled discharge valve is open, and the more powder is output. The adaptive control module has a pre-calibrated duty cycle-discharge quantity mapping table, which establishes precise mapping relationships for different powder types (due to differences in powder particle size, flowability, etc.) to ensure accurate control of the discharge quantity.

[0033] According to one embodiment of the present invention, it further includes: The wireless communication module is used to establish wireless communication connections with user smart terminals and cloud servers. The wireless communication module supports Bluetooth Low Energy (BLE) protocol and Wi-Fi protocol. In short-range scenarios, BLE connection is used first to reduce power consumption, and Wi-Fi connection is switched when a large amount of data needs to be transmitted. A user interaction module, located on the surface of the makeup powder compact, displays skin condition data, environmental parameter data, foundation formula recommendations, and makeup tutorial instructions. The user interaction module includes a touchscreen display using OLED (Organic Light Emitting Diode) or LCD (Liquid Crystal Display) technology, with a size ranging from 1.5 inches to 3.5 inches. The touchscreen display surface is coated with an anti-fingerprint and anti-glare coating to improve readability in bright outdoor light conditions. The data storage module is used to store historical skin condition data, historical environmental parameter data, historical foundation formula records, and user preference data. The data storage module uses non-volatile memory (such as Flash memory or EEPROM) with a storage capacity of not less than 4GB, supporting at least 6 months of data storage. The system also includes a power management module to provide power to all functional modules and supports both wireless charging and USB charging. The power management module comprises a rechargeable lithium battery (capacity 1500mAh to 3000mAh), a charge / discharge management circuit, a power monitoring circuit, and a low-power standby management circuit. When the low-power standby management circuit detects inactivity for more than a preset time (e.g., 5 minutes), it automatically puts all functional modules into sleep mode, retaining only the intermittent wake-up function of the wireless communication module to maximize battery life.

[0034] According to one embodiment of the present invention, the AI ​​analysis and processing module further includes a learning optimization unit, which is used to: collect user satisfaction feedback data for each generated foundation formula; construct a training sample set based on the satisfaction feedback data, combined with skin condition data, environmental parameter data, and actual foundation formula data of the corresponding batch; and use an incremental learning algorithm to update the parameters of the skin-environment-powder mapping model online, so that the subsequently generated foundation formula recommendation information gradually approaches the user's personalized preferences.

[0035] According to one embodiment of the present invention, the present invention also provides a multifunctional cosmetic powder box that further provides at least one of the following functional modules: The makeup simulation module is used to generate a virtual makeup effect image based on the skin condition data and the foundation formula recommendation information and display it on the user interaction module. The virtual makeup effect image can display the expected effect of the user's face after applying the recommended foundation formula in real time, helping the user to preview the makeup effect and make adjustments before mixing. The makeup tutorial guidance module is used to retrieve corresponding step-by-step makeup tutorials from the cloud tutorial database based on the skin condition data and the foundation formula recommendation information, and then play them through the user interaction module. The makeup tutorials include foundation application techniques, concealer skills, and setting methods, and support video playback and voice narration. The product lifecycle management module is used to generate a replenishment reminder when the remaining powder level is below a preset threshold, based on the monitoring results of the powder level monitoring unit, and send it to the user's smart terminal via the wireless communication module. The product lifecycle management module is also used to generate a replacement reminder when the powder is nearing its expiration date, based on the powder storage time and preset powder expiration information. The skin trend tracking module is used to record and analyze the changing trends of skin condition data over a long period, generate skin condition trend reports, and adjust foundation formula recommendation strategies based on these reports. The skin condition trend reports include weekly / monthly change curves of skin moisture content, changes in skin oil content, and shifts in skin tone, helping users understand the long-term patterns of their skin condition.

[0036] In a preferred embodiment of the present invention, the cosmetic powder compact body adopts a split design, including a main module and a detachable powder storage module. The main module integrates a skin detection module, an environmental sensing module, an AI analysis and processing module, an adaptive control module, a wireless communication module, a user interaction module, a data storage module, and a power management module. The powder storage module integrates a powder management module and is detachably connected to the main module via a magnetic or snap-on structure. Users can replace different types of powder storage modules as needed, such as modules containing different shades of base powder to adapt to seasonal changes or changes in skin tone.

[0037] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: Multifunctional and highly integrated. This invention is the first to organically integrate five major functional modules—skin detection, environmental perception, powder management, AI analysis and processing, and adaptive control—into a portable cosmetic powder box system, achieving a fully intelligent closed-loop process from skin condition perception, environmental factor analysis, personalized formula generation to automatic dispensing.

[0038] Precise skin condition detection. Through the collaborative work of a multispectral image sensor (including visible light, near-infrared and ultraviolet fluorescence imaging units) and a contact sensor array, it can comprehensively assess the user's skin condition from multiple dimensions such as skin color, moisture, oil, texture, and deep features, with detection accuracy significantly higher than single-sensor solutions.

[0039] Environmentally adaptive formula adjustment. An environmental sensing module collects real-time environmental parameters such as temperature, humidity, light intensity, UV radiation, and air quality. An AI analysis and processing module dynamically adjusts the foundation formula based on these environmental factors, ensuring optimal performance in various environments.

[0040] Personalized continuous optimization capabilities. By collecting user feedback data through learning optimization units, and employing incremental learning algorithms to continuously optimize the skin-environment-powder mapping model, the system can "understand the user better with use," and the formula recommendation results gradually approach the user's optimal personalized needs.

[0041] Precise powder mixing control. The electronically controlled discharge valve is driven by a PWM control signal. Combined with a pre-calibrated duty cycle-discharge mapping table, precise proportioning control of various basic powders is achieved, with discharge error controlled within ±0.05g.

[0042] The system also offers a wealth of value-added features, including makeup simulation, makeup tutorials, product lifecycle management, and skin trend tracking, providing users with a one-stop intelligent beauty experience.

[0043] A balance between portability and intelligence. The system adopts a compact system architecture design and highly integrated electronic components, achieving a high level of intelligent functions while maintaining portability (overall weight not exceeding 250g and dimensions not exceeding 100mm×80mm×30mm).

[0044] Modular and scalable design. Utilizing a split structure, the powder storage module is removable and replaceable, allowing users to flexibly select different combinations of powders as needed, and facilitating equipment maintenance and upgrades. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0046] Figure 2 This is a schematic diagram of the modular structure of a multifunctional cosmetic powder box system provided as an embodiment of the present invention.

[0047] Reference numerals: 100, powder compact body; 110, skin detection module; 111, visible light imaging unit; 112, near-infrared imaging unit; 113, ultraviolet fluorescence imaging unit; 114, capacitive moisture sensor; 115, photoelectric oil sensor; 120, environmental sensing module; 121, temperature sensor; 122, humidity sensor; 123, ambient light sensor; 124, ultraviolet sensor; 125, air quality sensor; 130, powder management module; 131, powder storage bin; 132, remaining powder level. Monitoring Unit; 133. Powder Mixing Bin; 134. Mixing Control Unit; 140. AI Analysis and Processing Module; 141. Data Preprocessing Unit; 142. Skin Condition Assessment Unit; 143. Formula Generation Unit; 144. Cloud Collaboration Unit; 145. Learning and Optimization Unit; 150. Adaptive Control Module; 151. Formula Analysis Unit; 152. Signal Generation Unit; 153. Data Feedback Unit; 160. Wireless Communication Module; 170. User Interaction Module; 180. Data Storage Module; 190. Power Management Module. Detailed Implementation

[0048] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0049] In the description of this application, it should be noted that the terms "upper," "lower," "inner," "outer," "top / bottom," etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0050] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installed," "equipped with," "sleeved / connected," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Example 1: Basic Structure of a Multifunctional Cosmetic Powder Box System

[0051] like Figure 1 and Figure 2 As shown, this embodiment provides a multifunctional cosmetic powder box system 100, including: A skin detection module 110, mounted on the powder compact, is used to collect data on the user's skin condition. In this embodiment, the skin detection module 110 includes a multispectral image sensor and a skin contact sensor group. The multispectral image sensor includes a visible light imaging unit 111, a near-infrared imaging unit 112, and an ultraviolet fluorescence imaging unit 113. The skin contact sensor group includes a capacitive moisture sensor 114 and a photoelectric oil sensor 115.

[0052] The visible light imaging unit 111 employs a CMOS image sensor with a resolution of 8 megapixels to acquire visible light images of the user's facial skin. After preprocessing, the acquired image data can be used to extract skin tone data (including the mean and variance of the L, a, and b channels in the L*a*b* color space) and skin texture data (including characteristic parameters such as texture roughness, orientation, and contrast).

[0053] The near-infrared imaging unit 112 employs an InGaAs image sensor, operating in the 900nm to 1700nm wavelength range, to acquire deep images of the user's facial skin. Near-infrared light can penetrate the epidermis to obtain structural information of the dermis, thereby extracting skin moisture distribution data and subcutaneous feature data, providing a basis for assessing the deep condition of the skin.

[0054] The ultraviolet fluorescence imaging unit 113 uses an ultraviolet LED as the excitation source (excitation wavelength of 365nm) and a CMOS image sensor to acquire fluorescence images of the skin surface. By analyzing the brightness and distribution of the fluorescence images, the distribution of bacteria and cosmetic residues on the skin surface can be detected, providing a quantitative basis for assessing skin cleanliness.

[0055] The capacitive moisture sensor 114 uses an interdigitated electrode structure to estimate the moisture content of the stratum corneum by measuring the change in capacitance when in contact with the skin. The measurement range is 0% to 100%, and the accuracy is ±3%.

[0056] The photoelectric oil sensor 115 adopts the principle of multi-angle light scattering measurement. It calculates the oil content on the skin surface by analyzing the intensity ratio of scattered light at different angles. The measurement range is from 0 μg / cm² to 200 μg / cm², and the accuracy is ±5 μg / cm².

[0057] An environmental sensing module 120 is mounted on the powder compact body and is used to collect environmental parameter data of the usage environment. In this embodiment, the environmental sensing module 120 includes a temperature sensor 121, a humidity sensor 122, an ambient light sensor 123, an ultraviolet sensor 124, and an air quality sensor 125.

[0058] Temperature sensor 121 is a digital temperature sensor with a measurement range of -10℃ to 60℃ and an accuracy of ±0.5℃. Humidity sensor 122 is a capacitive humidity sensor with a measurement range of 0%RH to 100%RH and an accuracy of ±3%RH. Ambient light sensor 123 is a photodiode sensor with a measurement range of 0.01 lux to 100,000 lux, used to assess current ambient light conditions. Ultraviolet sensor 124 uses a GaN-based ultraviolet photodetector to measure the radiation intensity of the UVA and UVB bands, with a measurement range of 0 to 15 UV index levels. Air quality sensor 125 is a laser scattering PM2.5 sensor used to detect the concentration of PM2.5 in the air.

[0059] The powder management module 130 is located inside the cosmetic powder compact and is used to manage the storage information and usage records of various foundation ingredients. In this embodiment, the powder management module 130 includes five powder storage compartments 131, a powder balance monitoring unit 132, a powder mixing compartment 133, and a mixing control unit 134.

[0060] The five powder storage compartments 131 store the following basic powders respectively: the first storage compartment stores light-colored basic powders (shade numbers corresponding to users with lighter skin tones); the second storage compartment stores medium-colored basic powders (shade numbers corresponding to natural skin tones); the third storage compartment stores dark-colored basic powders (shade numbers corresponding to users with darker skin tones); the fourth storage compartment stores functional powders with sun protection function (containing physical sunscreens titanium dioxide and zinc oxide); and the fifth storage compartment stores functional powders with moisturizing function (containing moisturizing ingredients such as hyaluronic acid).

[0061] Each powder storage bin 131 is equipped with a discharge port and an electrically controlled discharge valve. The electrically controlled discharge valve adopts a spiral conveying structure driven by a micro stepper motor. The discharge amount is precisely controlled by controlling the number of rotation steps of the stepper motor, and the discharge amount per step is about 0.01g.

[0062] The powder balance monitoring unit 132 uses a capacitive level sensor, which is installed on the inner wall of each powder storage bin. It detects the height of the powder by measuring the change in the dielectric constant inside the bin, and then converts it into a balance percentage.

[0063] The powder mixing chamber 133 is connected to the outlet of each powder storage chamber and is used to receive and mix the powder output from each powder storage chamber. The powder mixing chamber 133 is equipped with a micro vibration motor with a vibration frequency of 100Hz to 200Hz and an amplitude of 0.2mm to 0.5mm, which promotes the uniform mixing of different powders through high-frequency micro-vibration.

[0064] The mixing control unit 134 uses a 32-bit ARM Cortex-M4 microcontroller, which is electrically connected to the electrically controlled discharge valve and the powder balance monitoring unit 132. It is used to receive the allocation instructions from the adaptive control module 150 and control the operation of each electrically controlled discharge valve.

[0065] The AI ​​analysis and processing module 140 is electrically connected to the skin detection module 110, the environmental perception module 120, and the powder management module 130, respectively. In this embodiment, the AI ​​analysis and processing module 140 includes a data preprocessing unit 141, a skin condition assessment unit 142, a formula generation unit 143, a cloud collaboration unit 144, and a learning optimization unit 145.

[0066] The data preprocessing unit 141 performs normalization and feature extraction on the raw data collected by the skin detection module 110 and the environment perception module 120. The normalization process uses the min-max normalization method to map each dimension of the data to the [0, 1] interval. Feature extraction includes: extracting texture features from visible light images using the Local Binary Pattern (LBP) algorithm; extracting hue features using the color moment method; extracting moisture distribution features from near-infrared images using the Gray-Level Co-occurrence Matrix (GLCM); and performing temporal smoothing filtering on the environmental parameter data.

[0067] The skin condition assessment unit 142 determines the user's current skin condition level based on a multi-dimensional feature vector and a preset deep neural network classification model. The deep neural network structure used in this embodiment is as follows: input layer (32 dimensions), first fully connected layer (256 neurons, ReLU activation), second fully connected layer (128 neurons, ReLU activation), dropout layer (dropout probability 0.3), third fully connected layer (64 neurons, ReLU activation), and output layer (5 neurons, Softmax activation, corresponding to five skin conditions: dry, normal, oily, combination, and sensitive).

[0068] The formula generation unit 143 generates foundation formula recommendations based on the current skin condition level and environmental parameter data, combined with the available powder information provided by the powder management module 130. Formula generation employs a combination of a rule-based expert system and a machine learning model. First, the expert system determines the basic formula framework based on the skin condition level and environmental parameters; then, the machine learning model (using a gradient boosting decision tree algorithm) refines the basic formula based on the user's historical preference data, ultimately outputting formula recommendations that include the type identifiers and precise proportions of each powder component.

[0069] The cloud collaboration unit 144 interacts with the cloud server via a wireless communication module, receives updated skin condition classification model parameters and formula mapping parameters from the cloud, and uploads locally anonymized usage data to the cloud to provide data support for the global optimization of the cloud model.

[0070] The learning optimization unit 145 collects user satisfaction feedback data for each generated foundation formula (obtained through the user interaction module 170, including five-star ratings from users on four dimensions: color matching, concealing effect, makeup duration, and skin comfort). It then uses the Elastic Weight Consolidation (EWC) algorithm to update the parameters of the skin-environment-powder mapping model online. The core idea of ​​the EWC algorithm is to measure the importance of each model parameter using the Fisher information matrix, applying strong regularization constraints to important parameters when learning new data, thereby achieving continuous learning while avoiding catastrophic forgetting.

[0071] The adaptive control module 150 is electrically connected to the AI ​​analysis and processing module 140 and the powder management module 130, respectively. In this embodiment, the adaptive control module 150 includes a formula analysis unit 151, a signal generation unit 152, and a data feedback unit 153.

[0072] The formula analysis unit 151 receives foundation formula recommendation information and analyzes the type identifier and proportion parameters (expressed as a weight percentage) of each basic powder.

[0073] The signal generation unit 152 generates PWM control signals for the electrically controlled discharge valves of each powder storage silo based on the type identifier and proportioning parameters. The frequency of the PWM signal is set to 1kHz, and the duty cycle is calculated based on the target discharge amount and a pre-calibrated duty cycle-discharge amount mapping table. The calibration process is completed before shipment: for each type of powder, the actual discharge amount is measured at different duty cycles, an accurate mapping curve is established, and stored in non-volatile memory.

[0074] After performing the mixing operation, the data feedback unit 153 updates the powder usage record in the powder management module 130 and feeds back the usage data (including actual output, user feedback rating, etc.) to the learning and optimization unit 145 of the AI ​​analysis and processing module 140.

[0075] The wireless communication module 160 supports Bluetooth Low Energy (BLE) 5.0 and Wi-Fi 802.11n protocols, and is used to establish wireless communication connections with user smart terminals and cloud servers.

[0076] The user interaction module 170 features a 2.4-inch OLED touchscreen display with a resolution of 480×800 pixels, located on the surface of the makeup powder compact. The user interaction module 170 displays the following information: skin condition data (shown in numerical and graded charts), environmental parameter data (temperature, humidity, UV index, etc.), foundation formula recommendations (names and proportions of each powder component), and makeup tutorial guidance (step-by-step illustrated tutorials).

[0077] The data storage module 180 uses 8GB eMMC flash memory to store historical skin condition data, historical environmental parameter data, historical foundation formula records, and user preference data.

[0078] The power management module 190 includes a 2000mAh rechargeable lithium battery, a Type-C USB charging port, a wireless charging receiver coil (supporting the Qi wireless charging standard), as well as charge / discharge management circuitry and low-power standby management circuitry. The system automatically enters deep sleep mode after more than 5 minutes of standby, reducing power consumption to below 50μA, and can support approximately 30 days of normal use on a full charge. Example 2: System Workflow

[0079] This embodiment describes in detail the typical workflow of the multifunctional cosmetic powder box system 100, including the skin condition detection process, the formula generation process, the powder mixing process, and the feedback learning process.

[0080] (1) Skin condition detection process The user first wakes up the system through the user interaction module 170 and selects the "Start Detection" function. After the system starts, the user aligns their facial skin with the multispectral image sensor of the skin detection module 110, maintaining a distance of approximately 20cm. The system then displays a real-time view and provides location guidance prompts through the user interaction module 170.

[0081] Visible light imaging unit 111, near-infrared imaging unit 112, and ultraviolet fluorescence imaging unit 113 acquire image data sequentially or simultaneously. Simultaneously, the user briefly touches their facial skin to capacitive moisture sensor 114 and photoelectric oil sensor 115, completing contact detection after approximately 5 seconds.

[0082] The data preprocessing unit 141 normalizes the collected raw data and extracts multidimensional feature vectors. These multidimensional feature vectors include: skin tone features (mean and variance of L, a, and b channels, 6 dimensions in total), skin oil features (oil content and its gradient changes, 2 dimensions in total), skin moisture features (moisture content and its spatial distribution parameters, 4 dimensions in total), skin texture features (roughness, orientation, and contrast, 3 dimensions in total), deep skin features (subcutaneous structure parameters extracted from near-infrared images, 5 dimensions in total), ambient temperature features (1 dimension), ambient humidity features (1 dimension), ambient light features (1 dimension), ultraviolet radiation intensity features (1 dimension), and air quality features (PM2.5 concentration, 1 dimension). A total of 25 feature vectors are input into the skin condition assessment unit 142.

[0083] The skin condition assessment unit 142 inputs a multi-dimensional feature vector into a deep neural network classification model, outputs the confidence distribution of each skin condition level, and takes the level with the highest confidence as the current skin condition assessment result. The assessment result is displayed to the user through the user interaction module 170.

[0084] (2) Formula generation process After the skin condition assessment is completed, the formula generation unit 143 receives the current skin condition level and environmental parameter data, and starts the formula generation process.

[0085] First, the formula generation unit 143 retrieves a basic formula framework from a pre-set expert rule base based on the skin condition level. The basic formula framework defines the recommended proportion range of each basic powder ingredient. For example, for "dry skin," the proportion range of moisturizing powder ingredients in the basic formula framework is set to 20% to 35%, which is higher than the 5% to 15% for "oily skin."

[0086] Secondly, the formula generation unit 143 adjusts the basic formula framework based on environmental parameter data: if the ultraviolet intensity exceeds the preset threshold (e.g., UV index ≥ 6), the proportion of sunscreen powder is increased; if the PM2.5 concentration exceeds the preset threshold (e.g., ≥ 75 μg / m³), the proportion of antioxidant powder is increased; if the ambient humidity is below 40% RH, the proportion of moisturizing powder is further increased.

[0087] Next, the formula generation unit 143 calls a machine learning model (gradient boosting decision tree) to perform personalized fine-tuning of the formula based on the user's historical preference data (stored in the data storage module 180). The machine learning model takes skin condition features, environmental parameter features, and historical preference features as input and outputs the adjustment amount of the proportion of each powder ingredient.

[0088] Finally, the formula generation unit 143 outputs foundation formula recommendation information, including the type identification of each basic powder ingredient, precise proportioning parameters (with an accuracy of 0.1%), suggested mixing order, and usage instructions. The formula recommendation information is displayed through the user interaction module 170, where the user can choose to "confirm mixing", "manually adjust", or "regenerate".

[0089] (3) Powder preparation process After the user confirms the recommended formula, the adaptive control module 150 receives the foundation formula recommendation information and starts the powder mixing process.

[0090] The formula analysis unit 151 analyzes the powder type identifiers and proportion parameters in the formula recommendation information. The signal generation unit 152 calculates the target output amount based on the proportion parameters (the system defaults to a single batch total of 2g, which can be adjusted from 1g to 5g according to user needs), looks up the table to obtain the PWM duty cycle corresponding to each powder, and generates a PWM control signal to send to the powder management module 130.

[0091] After receiving the PWM control signal, the mixing control unit 134 sequentially controls the electrically controlled discharge valves of each powder storage bin. The discharge sequence follows the mixing sequence recommended in the formula: first, the base color powder with a larger proportion is output, then the functional powder is output, and finally, the fine-tuning color powder with a smaller proportion is output.

[0092] After each powder component enters the powder mixing chamber 133 in sequence, the micro vibration motor starts and vibrates at a frequency of 150Hz for about 15 seconds to ensure that the powder components are thoroughly mixed. After mixing is complete, the user interaction module 170 displays a "mixing complete" message, and the user can take out the mixed foundation from the outlet of the powder mixing chamber 133 for use.

[0093] The data feedback unit 153 records the actual output of this batch, updates the powder usage record, and sends the data to the AI ​​analysis and processing module 140 for archiving.

[0094] (4) Feedback Learning Process After using the prepared foundation, users can rate their satisfaction with the formula through the user interaction module 170. The rating dimensions include color matching, coverage, makeup lasting time, and skin comfort, with each dimension rated on a five-star scale from 1 to 5 stars.

[0095] The learning optimization unit 145 collects satisfaction score data and combines it with skin condition data, environmental parameter data, and actual formula data from the corresponding batches to construct training samples. The EWC algorithm is used to incrementally update the parameters of the skin-environment-powder mapping model.

[0096] The loss function of the EWC algorithm is defined as: Where: L(θ) represents the overall loss function; L new (θ) represents the loss function for the new task; λ represents the hyperparameter controlling the strength of memory for the old task. These are the diagonal elements of the Fisher information matrix, used to measure parameters. The importance of the old task; Indicates the current model parameters; This represents the model parameters after the old task has been learned.

[0097] Through the aforementioned incremental learning mechanism, the system can continuously accumulate user preference information, making the subsequently generated recipe recommendations increasingly meet the user's personalized needs. Example 3: Implementation of Value-Added Functions

[0098] This embodiment describes the value-added functions provided by the multifunctional cosmetic powder box system 100 and their implementation methods.

[0099] (1) Makeup simulation function The makeup simulation module is based on augmented reality (AR) technology, using the front-facing camera of the user interaction module 170 (integrated next to the OLED touchscreen) to capture real-time images of the user's face. The system employs a deep learning-based facial landmark detection algorithm (detecting 68 facial landmarks) to locate areas such as the forehead, cheeks, bridge of the nose, and chin on the user's facial image.

[0100] Based on the skin tone data obtained by the skin detection module 110 and the currently generated foundation formula recommendation information, the system uses an image fusion algorithm (based on Poisson image editing) to render a virtual foundation effect in the corresponding area of ​​the user's facial image, generate a virtual makeup effect image, and display it in real time on the user interaction module 170.

[0101] Users can adjust the transparency of the virtual foundation effect using a slider control to compare the difference between before and after makeup application. Users can also choose to "save the effect" or "share to social media platforms".

[0102] (2) Makeup tutorial guidance function The makeup tutorial guidance module retrieves matching step-by-step makeup tutorials from the cloud tutorial database based on the skin condition assessment results obtained by the skin detection module 110 and the currently generated foundation formula recommendation information.

[0103] The tutorial covers: pre-makeup base steps, foundation application techniques (including dotting, patting, and light blending), concealing techniques, and setting methods. The tutorial is presented in text, images, or short videos via the user interaction module 170, and supports voice narration. Users can control the playback progress using controls such as "previous step," "next step," and "replay."

[0104] The tutorial database contains content recorded and reviewed by professional makeup artists and updated regularly. The system also supports recommending tutorials tailored to a user's skin condition and skill level based on their usage data.

[0105] (3) Product lifecycle management function The product lifecycle management function module reads the monitoring data from the powder balance monitoring unit 132 in real time. When the balance of any powder storage bin is lower than the preset threshold (e.g., 20% remaining), the module displays a replenishment reminder through the user interaction module 170 and pushes the reminder to the user's smart terminal through the wireless communication module 160.

[0106] The functional module also generates a replacement reminder message when the powder is about to expire (e.g., less than 30 days away from the expiration date) based on the first use time of the powder in the powder storage bin (recorded by the data storage module 180) and the preset powder expiration information (the expiration date of basic color powder is 18 months and the expiration date of functional powder is 12 months).

[0107] Users can check the remaining quantity and expiration date of each powder through the smart terminal app, and jump to the online store with one click to make additional purchases.

[0108] (4) Skin trend tracking function The skin trend tracking module continuously records the skin condition data collected by the skin detection module 110 each time and stores it in the data storage module 180. The system automatically generates skin condition trend reports weekly and monthly, including: changes in skin moisture content, changes in skin oil content, changes in skin color shift, and changes in skin texture roughness.

[0109] The trend report is displayed in chart form through the user interaction module 170 and can be exported to the user's smart terminal App for detailed viewing via the wireless communication module 160.

[0110] The system also automatically adjusts its foundation formula recommendation strategy based on changes in skin condition. For example, if it detects a continuous decline in skin moisture content, the system will gradually increase the recommended proportion of moisturizing powder in the formula; if it detects a gradual darkening of skin tone (such as summer tanning), the system will adjust the proportion of base shade powder accordingly. Example 4: Cloud Collaboration and System Upgrade

[0111] This embodiment describes the cloud collaboration mechanism and system upgrade method of the multifunctional cosmetic powder box system 100.

[0112] (1) Cloud-based collaboration mechanism The cloud server deploys training and inference services for the global skin state classification model and the recipe mapping model. The cloud collaboration unit 144 establishes a secure data transmission channel with the cloud server via the wireless communication module 160 (using the TLS 1.3 encryption protocol).

[0113] Once the cloud server completes the training and validation of the new model (typically based on aggregated analysis of large amounts of anonymized user data), it pushes a model update notification to the device via the cloud collaboration unit 144. The device automatically downloads the update package while connected to Wi-Fi and charging, and completes the local replacement of model parameters the next time the system is idle.

[0114] The cloud collaboration unit 144 is also responsible for periodically uploading locally collected anonymized usage data (with user identification information removed) to the cloud server, providing data support for the global optimization of the cloud model. The uploaded data includes: skin condition detection results, environmental parameters, generated formula information, user satisfaction scores, etc. This data, after aggregation and analysis, is used to improve the generalization ability and recommendation accuracy of the global model.

[0115] (2) System firmware upgrade The system supports Over-the-Air (OTA) firmware upgrades. When a new firmware version is released, the cloud server pushes an upgrade notification to the device via the cloud collaboration unit 144. After user confirmation, the device automatically downloads the firmware package (approximately 20MB to 50MB) and completes installation and reboot while charging.

[0116] The firmware upgrade includes: updating algorithm model parameters, deploying code for new functional modules, optimizing system performance and stability, and fixing security vulnerabilities. OTA upgrades utilize differential upgrade technology, downloading and installing only the changed parts to save network bandwidth and upgrade time.

[0117] (3) Collaboration with smart terminal apps Users can install the accompanying app on a smart device (such as a smartphone) to connect to the multi-functional powder compact system 100 via Bluetooth or Wi-Fi. The app provides the following extended functions: Large-screen skin condition analysis report: View a more detailed skin analysis report on your mobile phone screen, including high-definition skin images, detailed interpretation of various indicators, and comparative analysis with peers; Remote formula customization: Manually adjust the foundation formula parameters on a mobile app and send the customized formula to the device for mixing via the wireless communication module 160; Beauty community interaction: Share your personal recipes and makeup effect pictures to the community, and browse other users' popular recipes and makeup tips; Intelligent reminder management: Set timed reminders for skin detection, powder replenishment, and skincare steps. Example 5: Optimal Range of System Parameters

[0118] This embodiment provides preferred ranges for key parameters of the multifunctional cosmetic powder box system 100, so that those skilled in the art can better implement the present invention.

[0119] (1) Sensor parameters The visible light imaging unit 111 preferably has an image resolution of 5 million to 12 million pixels, a frame rate of 15 fps to 30 fps, and a field of view of 60° to 90°.

[0120] The operating wavelength of the near-infrared imaging unit 112 is preferably from 850nm to 1550nm, and the image resolution is preferably from 1 million to 5 million pixels.

[0121] The measuring electrode area of ​​the capacitive moisture sensor 114 is preferably 10 mm² to 50 mm², and the measuring frequency is preferably 10 kHz to 100 kHz.

[0122] The measurement accuracy of temperature sensor 121 is preferably within ±0.3℃, and the measurement accuracy of humidity sensor 122 is preferably within ±2%RH.

[0123] (2) Powder mixing parameters The preferred range for a single batch mixing quantity is 1g to 10g, with an adjustable precision of 0.5g. The preferred frequency of the PWM control signal is 500Hz to 2kHz, with a duty cycle control precision of 1%. The preferred single-step discharge quantity of the electrically controlled discharge valve is 0.005g to 0.02g. The preferred vibration frequency of the powder mixing chamber 133 is 100Hz to 300Hz, and the preferred mixing time is 10 seconds to 30 seconds.

[0124] (3) System power consumption parameters Power consumption in normal operating mode is preferably controlled below 800mW, power consumption in sleep mode is preferably controlled below 1mW, and power consumption in deep sleep mode is preferably controlled below 0.5mW. The battery capacity is preferably between 1500mAh and 3000mAh, and the battery life under normal use when fully charged is no less than 7 days.

[0125] (4) System dimensional parameters The preferred dimensions of the powder compact are: length 80mm to 120mm, width 60mm to 90mm, and thickness 20mm to 35mm. The total weight (including battery and full powder) is preferably controlled within the range of 200g to 300g.

[0126] (5) AI model parameters The training data volume of the deep neural network classification model is preferably no less than 100,000 sets of labeled data, the number of training rounds is preferably 50 to 200, and the batch size is preferably 32 to 128. The hyperparameter λ of the EWC algorithm is preferably between 0.1 and 10, and is dynamically adjusted according to the similarity of the tasks.

[0127] The implementation principle of this invention is as follows: This invention discloses a multifunctional cosmetic powder box system, belonging to the field of intelligent beauty device technology. The system includes: a skin detection module 110, used to collect skin condition data such as skin tone, oil, moisture, and texture; an environmental sensing module 120, used to collect environmental parameter data such as temperature, humidity, light, and ultraviolet radiation; a powder management module 130, used to manage the storage and preparation of various foundation raw materials; an AI analysis and processing module 140, which generates personalized foundation formula recommendations based on skin condition data and environmental parameter data using a preset skin-environment-powder mapping model; and an adaptive control module 150, which controls the powder management module 130 to perform powder preparation operations according to the formula recommendation information. The system also integrates wireless communication, user interaction, data storage, and power management modules 190. This invention organically integrates skin detection, environmental sensing, AI analysis, and automatic preparation functions, realizing a fully intelligent closed loop from sensing to execution, and solving the technical problems of existing cosmetic powder boxes such as single function, lack of adaptability, and low degree of personalization.

[0128] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A multifunctional cosmetic powder compact, characterized in that, include: A skin detection module (110) is disposed on the powder compact body (100) and is used to collect the user's skin condition data. The skin detection module (110) includes at least one optical sensor and at least one moisture sensor. The skin condition data includes skin tone data, skin oil content data, skin moisture content data and skin texture data. An environmental sensing module (120) is installed on the powder compact body (100) and is used to collect environmental parameter data of the usage environment. The environmental parameter data includes ambient temperature, ambient humidity, ambient light intensity and ultraviolet index. The powder management module (130) is located inside the powder box body (100) and is used to manage the storage information and usage records of various foundation raw materials, including at least three different shades and functions of base powder. The AI ​​analysis and processing module (140) is electrically connected to the skin detection module (110), the environment perception module (120), and the powder management module (130), respectively, and is used to receive the skin state data and the environmental parameter data, perform analysis and processing based on the preset skin-environment-powder mapping model, and generate personalized foundation formula recommendation information. as well as An adaptive control module (150) is electrically connected to the AI ​​analysis and processing module (140) and the powder management module (130) respectively, and is used to receive the foundation formula recommendation information and control the powder management module (130) to perform corresponding powder mixing operations according to the foundation formula recommendation information.

2. A multifunctional cosmetic powder box according to claim 1, characterized in that, The skin detection module (110) includes a multispectral image sensor, which includes a visible light imaging unit (111) and a near-infrared imaging unit (112). The visible light imaging unit (111) is used to acquire visible light images of the user's facial skin to extract skin tone data and skin texture data, and the near-infrared imaging unit (112) is used to acquire deep images of the user's facial skin to extract skin moisture distribution data and subcutaneous feature data.

3. A multifunctional cosmetic powder box according to claim 1, characterized in that, The environmental sensing module (120) further includes an ultraviolet sensor (124) and an air quality sensor (125), and the environmental parameter data further includes ultraviolet intensity and PM2.5 concentration value; the AI ​​analysis and processing module (140) is further used to adjust the sun protection factor parameter and protective efficacy parameter in the foundation formula recommendation information according to the ultraviolet intensity and the PM2.5 concentration value.

4. A multifunctional cosmetic powder box according to claim 1, characterized in that, The AI ​​analysis and processing module (140) includes: The data preprocessing unit (141) is used to normalize and extract features from the skin state data and the environmental parameter data to generate a multidimensional feature vector; A skin condition assessment unit (142) is connected to the data preprocessing unit (141) and is used to determine the user's current skin condition level based on the multidimensional feature vector and a preset skin condition classification model. The formula generation unit (143) is connected to the skin condition assessment unit (142) and the powder management module (130) respectively, and is used to generate the foundation formula recommendation information based on the current skin condition level and the environmental parameter data, combined with the available powder information provided by the powder management module (130); and The cloud collaboration unit (144) is used to interact with the cloud server through the wireless communication module (160) and receive the updated skin condition classification model and formula mapping parameters sent from the cloud.

5. A multifunctional cosmetic powder box according to claim 4, characterized in that, The skin condition assessment unit (142) uses a skin condition classification model based on a deep neural network. The input layer of the deep neural network receives the multidimensional feature vector, and the output layer outputs the confidence distribution of the skin condition level. The multidimensional feature vector includes skin tone features, skin oil features, skin moisture features, skin texture features, ambient temperature features, ambient humidity features, and ambient light features.

6. A multifunctional cosmetic powder box according to claim 1, characterized in that, The powder management module (130) includes: At least three powder storage compartments (131) are used to store base powders of different color numbers or different functions, and each of the powder storage compartments (131) is provided with a discharge port and an electrically controlled discharge valve; A powder balance monitoring unit (132) is installed in each of the powder storage bins (131) to monitor the powder balance in each of the powder storage bins (131) in real time. A powder mixing chamber (133) is connected to the outlet of each of the powder storage chambers (131) and is used to receive and mix the powder output from each of the powder storage chambers (131); and The mixing control unit (134) is electrically connected to the electrically controlled discharge valve and the powder balance monitoring unit (132) and is used to receive the adjustment instructions from the adaptive control module (150) and control the opening duration and opening ratio of each of the electrically controlled discharge valves.

7. A multifunctional cosmetic powder box according to claim 1, characterized in that, The adaptive control module (150) is further used for: Receive the foundation formula recommendation information and parse the type identifier and ratio parameters of each basic powder contained in the foundation formula recommendation information; Based on the type identifier and proportioning parameters, pulse width modulation control signals are generated for the electrically controlled discharge valves of each powder storage bin (131) in the powder management module (130); After the mixing operation is performed, the powder usage record in the powder management module (130) is updated, and the usage data is fed back to the AI ​​analysis and processing module (140) for optimizing subsequent formula recommendations.

8. A multifunctional cosmetic powder box according to any one of claims 1 to 7, characterized in that, Also includes: The wireless communication module (160) is used to establish a wireless communication connection with the user's smart terminal and the cloud server; The user interaction module (170) is disposed on the surface of the makeup powder box body (100) and is used to display the skin condition data, the environmental parameter data, the foundation formula recommendation information and the makeup tutorial guidance information; The data storage module (180) is used to store historical skin condition data, historical environmental parameter data, historical foundation formula records, and user preference data; as well as The power management module (190) is used to provide working power for each functional module and supports both wireless charging and USB charging.

9. A multifunctional cosmetic powder box according to claim 8, characterized in that, The AI ​​analysis and processing module (140) further includes a learning optimization unit (145), which is used for: Collect user satisfaction feedback data for each generated foundation formula; Based on the aforementioned satisfaction feedback data, and combined with the skin condition data, environmental parameter data, and actual foundation formula data of the corresponding batches, a training sample set is constructed. An incremental learning algorithm is used to update the parameters of the skin-environment-powder mapping model online, so that the subsequently generated foundation formula recommendations gradually approach the user's personalized preferences.

10. A multifunctional cosmetic powder box according to claim 1, characterized in that, The multifunctional cosmetic powder box also provides at least one of the following functional modules: The makeup simulation function module is used to generate a virtual makeup effect image based on the skin condition data and the foundation formula recommendation information and display it in the user interaction module (170); The makeup tutorial guidance module is used to retrieve the corresponding step-by-step makeup tutorial based on the skin condition data and the foundation formula recommendation information, and play it through the user interaction module (170). The product lifecycle management function module is used to generate a supplementary reminder message when the powder balance is lower than a preset threshold based on the monitoring results of the powder balance monitoring unit (132), and send it to the user's smart terminal through the wireless communication module (160). The skin trend tracking module is used to record and analyze the changing trends of the skin condition data over a long period of time, generate a skin condition trend report, and adjust the foundation formula recommendation strategy based on the trend report.

Images