Skin condition evaluation method based on skin detection function head and intelligent beauty terminal

Abstract

The invention relates to the technical field of beauty intelligent terminals, multi-modal deep learning and transfer learning, image processing and the like, and provides a skin condition evaluation method based on a skin detection function head and a beauty intelligent terminal. And synchronous acquisition of visible and near-infrared multi-channel skin images and impedance spectrums is realized. According to the method, multi-modal data are aligned by taking a unified timestamp as a reference, dark current correction, wavelet denoising and environment standardization are introduced in a preprocessing stage, and a unified feature vector is constructed. The multi-modal cross attention model subjected to transfer learning fine tuning is fused with each modal feature and then outputs four indexes of water-oil balance, pigmentation, texture roughness and inflammation risk; and if a high-risk state is detected, subsequent physiotherapy is automatically limited, and a voice alarm is given out, so that a detection, evaluation and protection closed loop is formed, and the intelligent beauty safety and evaluation accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of beauty intelligent terminal, multimodal deep learning and transfer learning, image sensing, generation and enhancement, and in particular to a skin condition assessment method based on a skin detection functional head and an intelligent beauty terminal. Background Art

[0002] With the rapid development of the home beauty and skin care market, a large number of portable beauty terminals have emerged on the market. Most of these devices use a single sensing method: either analyzing skin color and blemishes through a visible light camera, or measuring skin moisture content with the help of bioelectrical impedance at a fixed frequency. Due to the single source of information and limited quantitative dimensions, users often need to use multiple devices alternately to obtain a relatively comprehensive skin condition reference. Not only is the operation cumbersome and costly, but the measurement conditions between different devices are difficult to maintain consistency, resulting in discrete results and difficulty in comparison. In addition, existing devices generally ignore the influence of environmental factors - the skin spectrum and electrical impedance characteristics will change significantly with temperature and humidity; if there is a lack of real-time environmental compensation, the evaluation results obtained by the same user in winter and summer may be completely opposite, which seriously weakens the credibility of the data. More importantly, most products are in a state of separation in the detection and treatment chain: the detection function and the treatment output are independent of each other, and the light energy or radio frequency power cannot be dynamically adjusted according to the real-time skin condition. When the user is in a high-risk state such as inflammation or allergy, high-power treatment may still be triggered, bringing the risk of secondary damage. On the other hand, traditional algorithms are mostly based on manual features or shallow models, which make it difficult to fully explore the coupling relationship between spectrum, electricity, and environment. The evaluation dimensions are limited and the generalization ability is poor. Once different skin colors or different lighting environments are used, the accuracy of the results will drop significantly.

[0003] In summary, the following technical problems exist in the prior art: a single sensing modality leads to insufficient dimensions for skin condition assessment, making it difficult to simultaneously quantify water, oil, pigment, texture and inflammation; multiple devices or multi-step operations increase the cost of use, making it difficult to synchronize measurement conditions and resulting in a lack of comparability of results; lack of environmental compensation and time synchronization causes data to be affected by temperature, humidity and instantaneous physiological fluctuations, resulting in poor stability; detection is separated from physical therapy, making it impossible to reduce power in real time for protection in high-risk scenarios; traditional algorithms lack depth, making it difficult to integrate cross-modal features, and the accuracy and adaptability of assessments need to be improved. Summary of the invention

[0004] In view of the deficiencies in the above-mentioned prior art, the present invention provides a skin condition assessment method and a beauty smart terminal based on a skin detection functional head to form a closed loop of detection, assessment and protection, thereby improving the safety of smart beauty and the accuracy of assessment.

[0005] In a first aspect, the present invention provides a skin condition assessment method based on a skin detection functional head, which runs on a main control module of a beauty terminal, and the skin condition assessment method comprises: The multispectral imaging module in the skin detection functional head is driven to collect skin reflection grayscale images frame by frame in a preset band range according to narrowband channels with a preset number of channels, and the obtained skin reflection grayscale images are stacked according to the timestamp T0 to generate a multispectral skin image with narrowband channels with a preset number of channels, and the electrical impedance analysis module is driven to apply a step-sweep frequency constant current signal with a preset signal frequency and peak current to the detection part within a preset window based on T0, and the AC voltage is synchronously measured and the real and imaginary impedance parts of each frequency point of the AC voltage are calculated using a phase-locked algorithm to obtain an impedance spectrum vector; After the temperature and humidity acquisition module in the skin detection functional head outputs the temperature and relative humidity of the skin contact point at a preset acquisition frequency, the temperature and humidity data with the timestamp closest to T0 are selected and associated with the data of the multispectral skin image and the data of the impedance spectrum vector, the data preprocessing unit is called to perform dark current correction and flat field correction on the multispectral image to obtain clean multispectral data, wavelet bandpass denoising is performed on the impedance spectrum vector to obtain clean impedance data, and the temperature and humidity data, the clean multispectral data and the clean impedance data are standardized and spliced ​​into a unified feature vector F; The feature vector F is input into a multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, and four indices including water-oil balance index, pigmentation index, texture roughness index and inflammation risk index are output; the four indices are obtained for display control and when it is identified that the inflammation risk index exceeds a preset threshold, the voice interaction module of the beauty terminal is triggered to broadcast an alarm and limit subsequent physical therapy output.

[0006] In a second aspect, the present invention provides a smart beauty terminal, comprising: A skin detection functional head, which is detachably mounted on a functional head connection seat at the front end of the handle assembly of the beauty terminal, wherein the skin detection functional head internally integrates a multi-spectral imaging module, an electrical impedance analysis module, and a temperature and humidity acquisition module; The main control module is arranged inside the handle assembly and runs the skin condition assessment method based on the skin detection functional head. The main control module supplies power to the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module through the electrical contacts in the functional head connection socket and completes data interaction.

[0007] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a skin condition assessment method based on a skin detection functional head and a beauty intelligent terminal. The method comprises the following steps: driving a multispectral imaging module in the skin detection functional head to collect a skin reflection grayscale image frame by frame in a preset band range according to a narrowband channel with a preset number of channels, stacking the obtained skin reflection grayscale image according to a timestamp T0 to generate a multispectral skin image with a narrowband channel with a preset number of channels, and driving the electrical impedance analysis module to apply a step-sweep frequency constant current signal with a preset signal frequency and a peak current to a detection part within a preset window based on T0, synchronously measuring an alternating current voltage and using a phase-locked algorithm to calculate the real part and the imaginary part of the impedance at each frequency point of the alternating voltage to obtain an impedance spectrum vector. After the temperature and humidity acquisition module in the skin detection functional head outputs the temperature and relative humidity of the skin contact point at a preset acquisition frequency, selecting the temperature and humidity data with a timestamp closest to T0 and correlating them with the data of the multispectral skin image and the data of the impedance spectrum vector, calling The data preprocessing unit performs dark current correction and flat field correction on the multispectral image to obtain clean multispectral data, performs wavelet bandpass denoising on the impedance spectrum vector to obtain clean impedance data, standardizes the temperature and humidity data, the clean multispectral data and the clean impedance data and splices them into a unified feature vector F, inputs the feature vector F into a multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, and outputs four indexes including a water-oil balance index, a pigmentation index, a texture roughness index and an inflammation risk index; obtains the four indexes for display control and, when it is identified that the inflammation risk index exceeds a preset threshold, triggers the voice interaction module of the beauty terminal to broadcast an alarm and limit subsequent physical therapy output, thereby realizing the synchronous acquisition of visible and near-infrared multi-channel skin images and impedance spectra, forming a closed loop of detection, evaluation and protection, and improving the safety and evaluation accuracy of smart beauty. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their description are used to explain the present invention and do not constitute an improper limitation of the present invention. Some specific embodiments of the present invention will be described in detail in an exemplary and non-restrictive manner with reference to the drawings. The same reference numerals in the drawings indicate the same or similar components or parts. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 It is a flow chart of a skin condition assessment method based on a skin detection functional head according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of the intelligent beauty terminal according to an embodiment of the present invention; Figure 3 The present invention is a schematic diagram of a state in which a skin detection function head and a handle component of the smart beauty terminal are assembled.

[0009] Description of reference numerals: 1. Handle assembly; 10. Touch screen; 2. Skin detection function head. DETAILED DESCRIPTION

[0010] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0011] Embodiment 1 See also Figure 1-Figure 3 This embodiment provides a skin condition assessment method based on a skin detection functional head, which runs on the main control module of the beauty terminal. The skin detection functional head 2 can be detachably installed on the functional head connection seat at the front end of the handle assembly 1 of the beauty terminal. The multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module are integrated inside the skin detection functional head; the main control module is arranged inside the handle assembly, and the main control module supplies power to the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module through the electrical contacts in the functional head connection seat and completes data interaction.

[0012] See also Figure 1-Figure 3 The skin condition assessment method based on the skin detection functional head provided in this embodiment includes the following steps: S101, driving the multispectral imaging module in the skin detection functional head to collect skin reflection grayscale images frame by frame in a preset band range according to narrowband channels with a preset number of channels, stacking the obtained skin reflection grayscale images according to timestamp T0 to generate a multispectral skin image with narrowband channels with a preset number of channels, and driving the electrical impedance analysis module to apply a step-sweep constant current signal with a preset signal frequency and peak current to the detection part within a preset window based on T0, synchronously measuring the AC voltage and using a phase-locked algorithm to calculate the real and imaginary impedance parts of each frequency point of the AC voltage to obtain an impedance spectrum vector; S102, after the temperature and humidity acquisition module in the skin detection functional head outputs the skin-tight point temperature TS and the relative humidity RHS at a preset acquisition frequency, the temperature and humidity data with the closest timestamp to T0 in the skin-tight point temperature TS and the relative humidity RHS are selected and associated with the data of the multispectral skin image and the data of the impedance spectrum vector, the data preprocessing unit is called to perform dark current correction and flat field correction on the multispectral image to obtain clean multispectral data, wavelet bandpass denoising is performed on the impedance spectrum vector to obtain clean impedance data, and the temperature and humidity data, the clean multispectral data and the clean impedance data are standardized and spliced ​​into a unified feature vector F; S103. Input the feature vector F into the multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, and output four indices including the water-oil balance index, the pigmentation index, the texture roughness index and the inflammation risk index; obtain the four indices for display control and when it is identified that the inflammation risk index exceeds a preset threshold, trigger the voice interaction module of the beauty terminal to broadcast an alarm and limit subsequent physical therapy output.

[0013] It should be noted that in this embodiment, a multispectral imaging module and an electrical impedance analysis module are deployed simultaneously in the same detection head. Multispectral images can reflect the differences in skin pigment, hemoglobin and collagen absorption in multiple narrowband channels in the visible and near-infrared, and the impedance spectrum can characterize the electrical properties such as the water content of the stratum corneum and the capacitance of the cell membrane; after the two complement each other, four types of physiological indicators, namely water, oil, pigment, texture and inflammation, can be inferred simultaneously in a single detection, getting rid of the dilemma of users going back and forth between multiple devices. In addition, in this embodiment, optical and electrical dual sampling is completed within the same time reference T0, and the sampling time difference is strictly limited by a preset window; then a frame of data with a timestamp closest to T0 is selected in the temperature and humidity acquisition module and bound to the above two types of data. Such a synchronization mechanism ensures that all modalities reflect the same physiological transient, improving the repeatability and credibility of cross-time comparison. In addition, in this embodiment, the synchronous temperature and humidity are included in the unified feature vector F as the third mode, and dark current and flat field correction are performed on the spectrum in the preprocessing stage, and wavelet denoising is performed on the impedance, and then unified and standardized, so as to eliminate the fixed noise of the sensor at the physical level, and introduce temperature and humidity deviation at the algorithm level, so that the subsequent model can automatically learn the coupling relationship between the environment and skin characteristics, and enhance regional versatility. In addition, in this embodiment, the inflammation risk index is output in real time after model reasoning, and the results exceeding the threshold are broadcast and warned, and the subsequent energy output is limited to form a closed loop of detection, evaluation and protection, and eliminate the hidden danger of secondary stimulation. Moreover, this embodiment adopts a multimodal cross-attention model fine-tuned by transfer learning. The model can automatically assign weights to spectral, impedance, and environmental features, and mine deep coupling information; transfer learning uses the prior knowledge pre-trained by large-scale public skin data sets to shorten the local annotation cycle and improve the adaptability to different skin colors and illumination.

[0014] It should also be noted that, in this embodiment, the preset band range may be 400nm-1000nm, the preset number of channels may be 16 narrowband channels; the preset window based on T0 may be a T0±5ms window; in the preset signal frequency and peak current, the signal frequency may be 0.1kHz-500kHz, and the peak current ≤200µA; the preset acquisition frequency may be 10Hz. Among them, 400nm-1000nm covers the absorption peaks of melanin, hemoglobin, and collagen; 16 channels ensure the balance between spectral resolution and hardware cost; 0.1kHz-500kHz sweep energy spectrum covers the change of conductive properties from extracellular fluid to intracellular fluid; 200µA peak current meets the limit of skin safety current; 10Hz temperature and humidity sampling frequency is sufficient to capture rapid fluctuations in indoor environment. The T0±5ms synchronization window ensures that the spectral frame and impedance sweep are in the same cardiac cycle to reduce errors.

[0015] See also Figure 3 , when obtaining the four indexes for display control, it may include: writing the four indexes together with the timestamp T0 into the data storage module of the beauty terminal, and controlling the four indexes stored in the data storage module to be displayed on the touch screen 10 of the beauty terminal in a card-type UI. Among them, the card-type UI instead of a list or a broken line can display the level, color and brief suggestions of the four indexes at one time on the limited handheld screen, lowering the reading threshold and shortening the decision-making time. The UI card can also carry the interactive logic of click jump, which is convenient for users to view historical curves or share to mobile phone APP.

[0016] In addition, when the feature vector F is input into the multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, the method includes: dividing the unified feature vector F into three parts, namely, spectral features, impedance features and environmental features, according to a predefined offset; using three groups of fully connected networks to map the three parts of features respectively to obtain a spectral embedding vector, an impedance embedding vector and an environmental embedding vector with the same size, and adding a sine-cosine position encoding based on the channel order after each embedding vector to retain the original order information; splicing the spectral embedding vector, the impedance embedding vector and the environmental embedding vector with the same size into a vector sequence in a predetermined order, and sending the spliced ​​vector sequence into a multi-layer cross-attention Transformer; In each layer of Transformer, the network calculates the attention weights between spectrum and impedance, spectrum and environment, and impedance and environment respectively, and dynamically fuses the mutual correlation information of the three modalities through residual connection and layer normalization; a CLS marker vector representing the overall information is set at the front end of the splicing sequence, and after being transmitted through the cross attention layer, the updated CLS vector is extracted as the fused global skin feature; the global skin feature is simultaneously input into four parallel regression branches, each branch consists of a two-level fully connected network and a Sigmoid activation function, and outputs the water-oil balance index, pigmentation index, texture roughness index and inflammation risk index respectively, and the value of each index is limited to between zero and one. In this embodiment, the multimodal cross attention model fine-tuned by transfer learning can fully explore the deep correlation between spectrum, impedance and environmental features in the same reasoning, and then give four skin health indexes at the same time, so as to provide a high-precision and real-time evaluation basis for the display, storage and safety power linkage of the beauty terminal.

[0017] In some preferred embodiments, the multi-spectral imaging module includes an EEPROM for storing spectral correction coefficients and a reference light source interface. The main control module first reads the spectral correction coefficients in the EEPROM and performs a CRC check after each power-on; if the check fails or the cumulative working time is detected to exceed a preset threshold, the main control module controls the integrating sphere reference light source built into the skin detection function head to output white field and dark field signals in sequence through the reference light source interface, calls the data preprocessing unit to complete the two-point calibration of the spectral correction coefficients and rewrites the updated spectral correction coefficients into the EEPROM, and at the same time displays the calibration progress bar on the touch screen of the beauty terminal until the calibration is completed before entering the frame-by-frame acquisition process of the skin reflection grayscale image. It should be noted that in this embodiment, an EEPROM storage spectral correction coefficient and an integrating sphere two-point calibration mechanism are introduced for the metrological stability of the multi-spectral channel drifting over time. After the main control module is powered on, the gain and bias matrices in the EEPROM are read and a CRC check is performed. If the correction coefficient is found to be damaged or the cumulative working time exceeds the threshold, the integrating sphere white field and dark field calibration are automatically called to correct the errors caused by LED aging, light path pollution or CMOS response attenuation in real time. The updated coefficients are written back to the EEPROM and the progress is displayed in real time in the UI to prevent users from mistaking the device for a malfunction. This can control the absolute error of the band reflectivity within 2%, solving the core problem of inaccuracy caused by the lack of self-calibration in household devices. Among them, the two-point calibration of the integrating sphere refers to: using the two extreme conditions of the reference white field (high reflection standard) and the reference dark field (zero light signal), by measuring the output response of the multi-spectral channel in the two states, establishing a channel-level gain and bias correction model, thereby completing the calibration of the spectral sensing system.

[0018] In some preferred embodiments, the electrical impedance analysis module uses a replaceable patch electrode. The main control module applies a 10kHz, 20µA detection pulse to the patch electrode and measures the open-circuit impedance before performing impedance acquisition; when the measured impedance value is higher than the preset aging threshold, the main control module stops the subsequent step frequency sweep process, pops up a prompt on the touch screen that the electrode needs to be replaced due to aging, and broadcasts the replacement prompt content through the voice interaction module. It should be noted that the interfacial electrochemical impedance of the replaceable patch electrode (such as silver-silver chloride electrode) increases after repeated use, which will cause high-frequency impedance spectrum distortion. The open-circuit impedance is measured by a low-amplitude test pulse and compared with the factory calibration threshold, so that the health status of the electrode can be determined before the formal frequency sweep. If the threshold is exceeded, the acquisition is suspended and the user is prompted to replace the electrode by voice and UI to avoid outputting an erroneous water-oil index or inflammation index due to electrode degradation.

[0019] In some preferred embodiments, when the temperature and humidity acquisition module detects that the temperature at the skin point is higher than 40°C or the relative humidity is lower than 20%, the main control module automatically reduces the LED duty cycle of the multispectral imaging module by 30%, shortens the exposure time of the CMOS sensor by 25%, and adds a set of environmental compensation vectors {ΔT, ΔRH} to the feature vector F, where ΔT and ΔRH are the deviation values ​​between the current temperature and humidity and the preset standard environment, so that the multimodal model skin state assessment model can dynamically correct the environmental impact during the inference stage. It should be noted that this embodiment proposes an integrated solution of dynamic power reduction, exposure adjustment, and environmental compensation vectors to address the dual interference of LED heating and skin dehydration on the measurement results in a high temperature and low humidity environment. First, at the physical level, the LED duty cycle is reduced by 30% and the exposure is shortened by 25% to avoid chip overheating and image saturation. Then, at the algorithm level, ΔT and ΔRH are injected into the feature vector F as additional dimensions, so that the multimodal skin condition assessment model can learn the temperature, humidity and skin response mapping during inference, thereby achieving domain adaptation, thereby preventing hardware damage and ensuring model accuracy, solving the problem of measurement drift caused by environmental changes, and ensuring that the pigment index and water-oil index are controlled within an explainable range when the temperature difference is high or low.

[0020] In some preferred embodiments, the multi-spectral imaging module adopts a variable aperture structure with a stepper motor during the acquisition process. The main control module calculates the target exposure value in real time according to the external illumination detected by the ambient light sensitive unit of the beauty terminal, and adjusts the blade angle of the variable aperture through a closed-loop PID control stepper motor, so that the average grayscale of the skin reflection grayscale image of the preset channel number is maintained between 128±5 to prevent spectral distortion caused by overexposure or underexposure; when the rate of change of the external illumination exceeds the threshold, the main control module temporarily suspends the frame-by-frame acquisition process of the skin reflection grayscale image until the PID output is stable. It should be noted that in this embodiment, the exposure is controlled in real time by the variable aperture and PID closed loop, and the mean of the 16-channel grayscale histogram can be locked at 128±5 to ensure the stability of the spectral signal-to-noise ratio. The ambient light sensitive unit feeds back the illumination in real time, and the PID controls the stepper motor to adjust the blade angle to suppress overshoot. If the sudden change of illumination exceeds the threshold, the acquisition is temporarily suspended to ensure that the generated multi-spectral image has no serious exposure drift, thereby improving the stability of the pigmentation index and the texture roughness index.

[0021] In some preferred embodiments, all data frames from the multi-spectral imaging module, the electrical impedance analysis module, and the temperature and humidity acquisition module are stamped with a 32-bit hardware timestamp before waiting for the splicing of the feature vector F. The main control module uses a maximum frame alignment tolerance of 1ms. If the data delay of any module in the multi-spectral imaging module, the electrical impedance analysis module, and the temperature and humidity acquisition module exceeds the tolerance, the corresponding frame is discarded and the corresponding module acquisition is re-triggered until the difference in the timestamps of the three types of data of the multi-spectral imaging module, the electrical impedance analysis module, and the temperature and humidity acquisition module meets the tolerance before the splicing of the feature vector F is performed. It should be noted that in this embodiment, a 32-bit hardware timestamp and a 1ms frame alignment tolerance are used to ensure that the time domain of the multi-modal data is consistent. When the main control module takes a frame, if the maximum and minimum difference is greater than 1ms, the frame is discarded and re-collected to ensure that the spectral frame, the impedance spectrum, and the temperature and humidity data reflect the same cardiac cycle and the same respiratory phase, and avoid the spectrum looking at 0.5 seconds ago and the impedance looking at the present, resulting in the multi-modal model skin state assessment model discrimination distortion.

[0022] In some preferred embodiments, the multimodal model skin state assessment model supports incremental updates of federated learning. After the user authorizes, the main control module regularly updates the local gradient ΔW for differential privacy processing and uploads it to the federated server. The federated server completes the aggregation to obtain the global model parameter W and pushes it back to the main control module via OTA. The main control module verifies the model signature and replaces the old weight and records the version number; if the signature verification fails, it rolls back to the previous stable version and reports the error log. It should be noted that in this embodiment, ΔW (local gradient) is uploaded with noise to protect the user's original data from leaving the device; after the server aggregates, OTA sends the new weight and can be verified with RSA signature to ensure that the model has not been tampered with. This embodiment enables the model to continuously iterate with the diversity of skin color, age, and environment of global users, improves generalization ability, and solves the problem of poor adaptability of shallow algorithms. Among them, OTA refers to over-the-air download. OTA can remotely download and update the software or model parameters on the device through wireless networks (such as Wi-Fi, Bluetooth, NFC, etc.) without the user manually disassembling the machine or connecting to a computer for upgrading.

[0023] In some preferred embodiments, the main control module also communicates with the intelligent power management unit, which monitors the battery power, temperature and internal resistance in real time. When it is detected that the remaining power is less than 15% or the battery temperature is higher than 45°C, the main control module reduces the LED output power of the multi-spectral imaging module by 50%, shortens the sweep frequency step length and limits the brightness of the touch screen to 60%, and displays a low power or high temperature warning bar at the top of the UI; if the temperature continues to rise to 55°C, the main control module enters sleep mode. It should be noted that in this embodiment, the LED power reduction, sweep frequency reduction, and screen brightness reduction are triggered by the remaining power <15% or the battery temperature >45°C, and the temperature is directly hibernated when it is >55°C, which reduces instantaneous power consumption and heat generation while ensuring the integrity of the detection, and prevents the voltage drop caused by the increase in the internal resistance of the battery from affecting the LED spectrum stability or impedance constant current accuracy. The UI warning bar prompts the user to charge or cool in real time, further improving the usability and safety of the beauty terminal in daily high-frequency use scenarios.

[0024] In some preferred embodiments, the main control module calls the posture sensing unit to detect the inclination of the skin detection function head relative to the direction of gravity before each detection process begins. When the inclination is greater than the inclination threshold or a sudden change in acceleration exceeding the acceleration threshold is detected, the main control module suspends the lighting of the LED of the multispectral imaging module, and prompts the user of abnormal fitting angle through the touch display screen and the voice interaction module. It should be noted that before detection, the posture sensing unit is used to determine that the inclination is greater than the inclination threshold, such as the inclination>30° or the acceleration>acceleration threshold, that is, the LED lighting is suspended and the user is prompted to correct the posture, thereby solving the problem of oblique light spot and poor contact of impedance electrodes caused by excessive user holding angles, and also preventing high-power LED from directly hitting the eyes when falling. The acceleration threshold can be based on the empirical value of the drop test, and the inclination threshold can be based on the results of optical path geometry simulation.

[0025] In some preferred embodiments, when the inflammation risk index exceeds the preset threshold and three consecutive tests are all high risk, the main control module automatically generates a comprehensive report file containing the detection trends, environmental parameters and user-reported symptoms in the past seven days, and provides a QR code download interface through the touch display screen, and embeds a SHA-256-based integrity check code in the comprehensive report file. It should be noted that when the inflammation risk index exceeds the threshold three times in a row, a seven-day comprehensive report is automatically generated and a SHA-256 check code is embedded, which is available for download through a QR code. Among them, three consecutive judgments can filter out occasional false alarms; the seven-day window covers a round of skin barrier renewal cycle, which is convenient for evaluating the efficacy. The report contains environmental data and user self-reports, plus SHA-256 verification, which can ensure that the report has not been modified and ensure the objectivity of the evaluation.

[0026] In some preferred embodiments, the multispectral imaging module supports an optional extended band adapter, and the extended adapter has a built-in 1000nm-1700nm InGaAs near-infrared detector array and a corresponding filter group; when the main control module detects that the extended adapter is inserted and reads the legal ID through I²C, the band scanning table is expanded to 24 channels, and a near-infrared moisture absorption peak correction algorithm is added to the data preprocessing unit, and the dimension mapping table of the feature vector F is updated at the same time. It should be noted that in this embodiment, the band is extended to 1700nm through a pluggable InGaAs near-infrared extended detector, which can detect the second-order absorption peak of moisture and the collagen scattering peak, thereby improving the accuracy of deep moisture and collagen assessment. I²C reads the legal ID to prevent unofficial adapters from damaging the host; the 24-channel scanning table and the near-infrared moisture correction algorithm are sent down at the same time to ensure consistent data dimensions.

[0027] In some preferred embodiments, the main control module reserves an OTA update interface and provides a firmware upgrade page in the mobile terminal APP. The firmware package is signed with RSA-2048 and transmitted through two-way TLS verification. During the upgrade process, the main control module first writes the new firmware to the standby partition and performs a CRC32 check. After the check passes, the boot partition is switched and restarted. If the startup fails, it automatically rolls back to the old version and records the abnormal reason. The upgrade log is sent to the APP via BLE broadcast for users to query. It should be noted that in this embodiment, when writing the firmware, the standby partition is written first, and the boot partition is switched after the CRC32 check passes, which can prevent the upgrade interruption from causing device failure; if it fails, it automatically rolls back and broadcasts the log via BLE, so that users can know the status in time.

[0028] Embodiment 2 See also Figure 1-Figure 3 This embodiment provides a smart beauty terminal, including: A skin detection functional head, which is detachably mounted on a functional head connection seat at the front end of the handle assembly of the beauty terminal, wherein the skin detection functional head internally integrates a multi-spectral imaging module, an electrical impedance analysis module, and a temperature and humidity acquisition module; The main control module is arranged inside the handle assembly and runs the skin condition assessment method based on the skin detection functional head. The main control module supplies power to the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module through the electrical contacts in the functional head connection socket and completes data interaction. The skin condition evaluation method drives the multispectral imaging module in the skin detection functional head to collect the skin reflection grayscale image frame by frame in the preset band range according to the narrowband channel with a preset number of channels, stacks the obtained skin reflection grayscale image according to the timestamp T0 to generate a multispectral skin image with a narrowband channel with a preset number of channels, and drives the electrical impedance analysis module to apply a step-sweep constant current signal with a preset signal frequency and peak current to the detection part within a preset window based on T0, synchronously measures the AC voltage and uses a phase-locked algorithm to calculate the real part and the imaginary part of the impedance of each frequency point of the AC voltage to obtain an impedance spectrum vector. After the temperature and humidity acquisition module in the skin detection functional head outputs the temperature and relative humidity of the skin contact point at the preset acquisition frequency, the temperature and humidity data with the timestamp closest to T0 are selected and associated with the data of the multispectral skin image and the data of the impedance spectrum vector, and the data preprocessing unit is called to process the multispectral skin image. The spectrum image is subjected to dark current correction and flat field correction to obtain clean multispectral data, wavelet bandpass denoising is performed on the impedance spectrum vector to obtain clean impedance data, the temperature and humidity data, the clean multispectral data and the clean impedance data are standardized and spliced ​​into a unified feature vector F, the feature vector F is input into a multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, and four indexes including water-oil balance index, pigmentation index, texture roughness index and inflammation risk index are output; the four indexes are obtained for display control and when it is identified that the inflammation risk index exceeds a preset threshold, the voice interaction module of the beauty terminal is triggered to broadcast an alarm and limit subsequent physical therapy output, thereby realizing the synchronous acquisition of visible and near-infrared multi-channel skin images and impedance spectra, forming a closed loop of detection, evaluation and protection, and improving the safety and evaluation accuracy of smart beauty.

[0029] It should be pointed out that the above embodiments are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A skin condition assessment method based on a skin detection function head, running on a main control module of a beauty terminal, characterized in that: The skin condition assessment method comprises: The multispectral imaging module in the skin detection functional head is driven to collect skin reflection grayscale images frame by frame in a preset band range according to narrowband channels with a preset number of channels, and the obtained skin reflection grayscale images are stacked according to the timestamp T0 to generate a multispectral skin image with narrowband channels with a preset number of channels, and the electrical impedance analysis module is driven to apply a step-sweep frequency constant current signal with a preset signal frequency and peak current to the detection part within a preset window based on T0, and the AC voltage is synchronously measured and the real and imaginary impedance parts of each frequency point of the AC voltage are calculated using a phase-locked algorithm to obtain an impedance spectrum vector; After the temperature and humidity acquisition module in the skin detection functional head outputs the temperature and relative humidity of the skin contact point at a preset acquisition frequency, the temperature and humidity data with the timestamp closest to T0 are selected and associated with the data of the multispectral skin image and the data of the impedance spectrum vector, the data preprocessing unit is called to perform dark current correction and flat field correction on the multispectral image to obtain clean multispectral data, wavelet bandpass denoising is performed on the impedance spectrum vector to obtain clean impedance data, and the temperature and humidity data, the clean multispectral data and the clean impedance data are standardized and spliced ​​into a unified feature vector F; The feature vector F is input into a multimodal cross-attention model fine-tuned by transfer learning to fuse the features of each modality, and four indices including water-oil balance index, pigmentation index, texture roughness index and inflammation risk index are output; the four indices are obtained for display control and when it is identified that the inflammation risk index exceeds a preset threshold, the voice interaction module of the beauty terminal is triggered to broadcast an alarm and limit subsequent physical therapy output.

2. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: When the four indexes are obtained for display control, it includes: writing the four indexes together with the timestamp T0 into the data storage module of the beauty terminal, and controlling the four indexes stored in the data storage module to be displayed on the touch display screen of the beauty terminal in a card-type UI.

3. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: The preset band range is 400nm-1000nm, the preset number of channels is 16 narrowband channels; the preset window based on T0 is a T0±5ms window; among the preset signal frequency and peak current, the signal frequency is 0.1kHz-500kHz, and the peak current is ≤200µA; the preset acquisition frequency is 10Hz.

4. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: The multi-spectral imaging module includes an EEPROM for storing spectral correction coefficients and a reference light source interface. The main control module first reads the spectral correction coefficients in the EEPROM and performs a CRC check each time it is powered on; if the check fails or it is detected that the cumulative working time exceeds a preset threshold, the main control module controls the integrating sphere reference light source built into the skin detection function head to output white field and dark field signals in sequence through the reference light source interface, calls the data preprocessing unit to complete the two-point calibration of the spectral correction coefficients and rewrites the updated spectral correction coefficients into the EEPROM, and at the same time displays a calibration progress bar on the touch screen of the beauty terminal until the calibration is completed before entering the frame-by-frame acquisition process of the skin reflectance grayscale image.

5. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: The skin detection functional head can be detachably mounted on the functional head connecting socket at the front end of the handle assembly of the beauty terminal, and the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module are integrated inside the skin detection functional head; the main control module is arranged inside the handle assembly, and the main control module supplies power to the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module through the electrical contacts in the functional head connecting socket and completes data interaction.

6. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: The electrical impedance analysis module adopts a replaceable patch electrode. Before performing impedance acquisition, the main control module applies a 10kHz, 20µA detection pulse to the patch electrode and measures the open-circuit impedance; when the measured impedance value is higher than the preset aging threshold, the main control module stops the subsequent step frequency scanning process, pops up a prompt on the touch display screen that the electrode is aged and needs to be replaced, and broadcasts the replacement prompt content through the voice interaction module.

7. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: When the temperature and humidity acquisition module detects that the temperature at the skin-contacting point is higher than 40°C or the relative humidity is lower than 20%, the main control module automatically reduces the LED duty cycle of the multispectral imaging module by 30%, shortens the exposure time of the CMOS sensor by 25%, and adds a set of environmental compensation vectors {ΔT, ΔRH} to the feature vector F, where ΔT and ΔRH are the deviation values ​​between the current temperature and humidity and the preset standard environment, respectively, so that the multimodal model skin condition assessment model can dynamically correct the environmental impact during the inference stage.

8. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: The multispectral imaging module adopts a variable aperture structure with a stepper motor during the acquisition process. The main control module calculates the target exposure value in real time according to the external illumination detected by the ambient light sensitive unit of the beauty terminal, and adjusts the blade angle of the variable aperture through a closed-loop PID control stepper motor, so that the average grayscale of the skin reflection grayscale image of a preset number of channels is maintained between 128±5, thereby preventing spectral distortion caused by overexposure or underexposure. When the rate of change of the external illumination exceeds a threshold, the main control module temporarily suspends the frame-by-frame acquisition process of the skin reflection grayscale image until the PID output is stable.

9. The skin condition assessment method based on the skin detection functional head according to claim 1, characterized in that: All data frames from the multispectral imaging module, electrical impedance analysis module and temperature and humidity acquisition module are marked with 32-bit hardware timestamps before waiting for the splicing of the feature vector F. The main control module adopts a maximum frame alignment tolerance of 1ms. If the data delay of any module among the multispectral imaging module, electrical impedance analysis module and temperature and humidity acquisition module exceeds the tolerance, the corresponding frame will be discarded and the corresponding module acquisition will be re-triggered until the timestamp difference of the three types of data from the multispectral imaging module, electrical impedance analysis module and temperature and humidity acquisition module meets the tolerance before the splicing of the feature vector F is performed.

10. An intelligent beauty terminal, characterized in that: include: A skin detection functional head, which is detachably mounted on a functional head connection seat at the front end of the handle assembly of the beauty terminal, wherein the skin detection functional head internally integrates a multi-spectral imaging module, an electrical impedance analysis module, and a temperature and humidity acquisition module; A main control module is arranged inside the handle assembly, and runs the skin condition assessment method based on the skin detection functional head as described in any one of claims 1 to 9. The main control module supplies power to the multi-spectral imaging module, the electrical impedance analysis module and the temperature and humidity acquisition module through the electrical contacts in the functional head connecting socket and completes data interaction.

Images