Laser semiconductor array spectrum matching optimization method and system fused with AI large model

Abstract

The invention provides a laser semiconductor array spectrum matching optimization method and system fused with an AI large model, and relates to the technical field of semiconductor laser array spectrum optimization. The method comprises the following steps: firstly, obtaining personalized medical beauty function requirements of a user, determining spectrum requirement information, and obtaining working parameter information according to a preset laser semiconductor array matching strategy library; acquiring output spectrum information in real time by utilizing spectrum detection equipment; determining parameter adjustment information through a spectrum matching optimization model constructed by deep learning in combination with spectrum requirements and output spectrum information; and finally adjusting the working parameters of the laser semiconductor array according to the parameter adjustment information. By implementing the method, the matching precision of the output spectrum of the laser semiconductor array and the personalized demand of the user can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of semiconductor laser array spectrum optimization, and in particular to a laser semiconductor array spectrum matching optimization method and system integrating an AI large model. Background Art

[0002] In the field of laser therapy, laser semiconductor arrays, as key light source components, have a performance that directly affects the effectiveness and safety of treatment. With the rapid development of medical technology, laser treatment methods have become increasingly diverse and sophisticated, from pigment spot removal and acne treatment in dermatology, to vision correction and fundus disease treatment in ophthalmology, to precise cutting and tissue repair in surgery, etc., which place extremely stringent requirements on laser parameters such as wavelength, power, and spectral purity. Different treatment scenarios and diseased tissues require lasers with specific wavelengths and spectral characteristics. For example, in dermatological treatment, tattoo removal requires a high-energy laser of a specific wavelength to crush pigment particles, while the treatment of port-wine stains requires a laser of another wavelength range to precisely act on vascular tissue. Therefore, achieving precise matching of the laser semiconductor array spectrum with various treatment needs has become a core requirement for improving the effectiveness of laser treatment and expanding the scope of treatment applications.

[0003] Currently, most systems rely on pre-set, fixed spectral parameter combinations. Before operation, technicians combine the operating parameters of different laser diode arrays based on common aesthetic medical needs and store them in the device. In actual use, the corresponding preset parameter combination is directly called based on the general aesthetic medical project type to control the laser diode array's output spectrum. This approach can meet some common aesthetic medical needs and achieve a certain degree of preliminary spectral matching.

[0004] However, with the development of the medical aesthetics industry, users' demand for personalized aesthetic results has become increasingly prominent. Existing spectral matching methods, based on fixed preset parameter combinations, are unable to adapt to the differences in skin type, age, and specific beauty preferences among different users. Because these fixed parameter combinations are difficult to flexibly adjust to each user's unique situation and needs, this can lead to deviations between spectral output and actual needs, compromising aesthetic results and even causing adverse effects on the skin due to inappropriate spectrum. Summary of the Invention

[0005] The present application provides a laser semiconductor array spectral matching optimization method and system integrated with an AI large model, which is used to solve the problems of low spectral matching accuracy and insufficient personalization in existing laser medical aesthetic treatments.

[0006] In the first aspect, the present application provides a laser semiconductor array spectral matching optimization method that integrates an AI large model, which is applied to a laser semiconductor array spectral matching optimization system. The method includes: obtaining the user's personalized medical beauty function demand information; determining the spectral demand information in combination with the personalized medical beauty function demand information; determining the working parameter information of the laser semiconductor array through a preset laser semiconductor array matching strategy library in combination with the spectral demand information, and the laser semiconductor array matching strategy library is constructed in advance based on multiple spectral demand information sets annotated with laser semiconductor array working parameter information; obtaining the output spectral information of the laser semiconductor array in real time through a spectral detection device; determining the parameter adjustment information of the laser semiconductor array through a spectral matching optimization model in combination with the spectral demand information and the output spectral information, and the spectral matching optimization model is constructed in advance based on multiple spectral demand information and output spectral information sets annotated with laser semiconductor array parameter adjustment information through deep learning; adjusting the working parameters of the laser semiconductor array in combination with the parameter adjustment information.

[0007] By adopting the above-mentioned technical solution, this method takes the user's personalized medical aesthetic functional needs as the starting point, uses large-scale model analysis to determine spectral requirements, then uses a preset strategy library to match the operating parameters of the laser semiconductor array. At the same time, it combines real-time spectral detection with optimization models to dynamically adjust parameters. This series of technical features interacts with each other to first accurately identify user needs, and then dually guarantee parameter accuracy based on historical data and real-time feedback. By real-time detection of the output spectrum and comparing it with the requirements, deviations can be promptly identified and adjusted through optimization model calculations, achieving closed-loop control. The ultimate effect is to significantly improve the matching accuracy of the laser semiconductor array output spectrum with the user's personalized needs, providing more accurate energy output for subsequent medical aesthetic treatments, and fundamentally improving treatment effectiveness.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of determining spectral demand information in combination with the personalized medical aesthetic function demand information specifically includes: inputting the personalized medical aesthetic function demand information into a pre-built large language model to determine the medical aesthetic function terms and modifiers in the personalized medical aesthetic function demand information, and the modifiers include at least part qualifiers and degree qualifiers; in combination with the medical aesthetic function terms, determining the basic spectral parameter information set corresponding to the medical aesthetic function term based on the spectral medical aesthetic function map; performing parameter correction on the basic spectral parameter information set according to the modifiers to generate spectral demand information.

[0009] By employing this technical solution, when determining spectral requirement information, a large language model is first used to analyze the medical aesthetic functional terms and modifiers in the user's requirements. The atlas then combines this with the baseline parameters to determine and modify them based on the modifiers. The large language model's precise parsing capabilities ensure accurate extraction of user intent. The atlas provides the baseline parameters, while the modifiers enable personalized adjustments. Site qualifiers adjust wavelength and intensity based on the tissue structure parameters of the target area, while degree qualifiers adjust pulse frequency and energy density based on the treatment depth level. This layered adjustment mechanism ensures that basic requirements are met while enabling refined personalized adaptation. The ultimate result is that the generated spectral requirement information is highly tailored to the user's specific needs, improving the targeted and effective treatment at the spectral parameter level.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the step of performing parameter correction on the basic spectral parameter information set according to the modifier to generate spectral demand information specifically includes: if the modifier includes a site qualifier, extracting tissue structure parameters corresponding to the site qualifier from a preset site optical property database, including epidermal thickness, dermal depth, and local pigment density; calculating the optical attenuation coefficient of the target site based on the tissue structure parameters; performing site adaptive adjustment on the spectral wavelength in the basic spectral parameter information set according to the optical attenuation coefficient, and at the same time, adjusting the optical attenuation coefficient of the basic spectral parameter information set according to the biological safety threshold of the target site. The spectrum intensity is adjusted accordingly; the adjusted spectrum wavelength and spectrum intensity are integrated to generate spectrum demand information adapted to the target site; if the modifier contains a degree qualifier, the degree qualifier is mapped to a predefined treatment depth level; according to the treatment depth level, the corresponding parameter adjustment rule is extracted from a preset depth adjustment strategy library; according to the parameter adjustment rule, the pulse frequency in the basic spectrum parameter information set is adjusted; based on the corrected pulse frequency and target depth, the corresponding energy input is calculated, and the energy density in the basic spectrum parameter information set is adjusted accordingly; the adjusted pulse frequency and energy density are integrated to generate spectrum demand information adapted to the corresponding treatment degree.

[0011] By adopting the above technical solutions, detailed adjustment strategies are designed for site qualifiers and degree qualifiers respectively. For site qualifiers, the optical attenuation coefficient is calculated taking into account tissue structural parameters such as epidermal thickness, and then the wavelength and intensity are adjusted to ensure that the spectrum can effectively reach the target tissue and meet the safety threshold. For degree qualifiers, they are mapped to the treatment depth level, and the pulse frequency and energy density are adjusted to achieve precise control of the treatment degree. These two adjustment strategies work together to optimize the basic spectral parameters from different dimensions. The ultimate effect is that the generated spectral demand information can adapt to the target site and treatment degree to the greatest extent, improve the utilization rate of light energy, reduce unnecessary damage to surrounding tissues, and at the same time ensure the stability and predictability of the treatment effect.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the step of adjusting the working parameters of the laser semiconductor array in combination with the parameter adjustment information, it also includes: real-time monitoring of status information of the patient's treatment site through a biosensor, where the status information includes at least tissue temperature; obtaining the patient's facial expression image data; combining the facial expression image data, determining the patient's adverse feedback data through a micro-expression feedback model, where the micro-expression feedback model is constructed in advance through deep learning based on multiple micro-expression image sets that are annotated with corresponding adverse feedback data under different treatment situations; combining the status information and the adverse feedback data to determine whether the current laser treatment has a negative impact on the patient; if a negative impact is generated, issuing an adverse impact prompt to the user end, and displaying the status information and adverse feedback data to the visual end.

[0013] By implementing the above technical solution, after adjusting the operating parameters of the laser semiconductor array, a real-time monitoring and feedback mechanism was added. Biosensors monitor status information such as tissue temperature, while facial expression image data is analyzed using a micro-expression feedback model to analyze patients' adverse reactions. These two data sources complement each other: status information reflects physiological changes, while adverse reaction data reflects the patient's subjective experience. By comprehensively analyzing these two data sources, potential negative effects during treatment can be promptly identified. If negative effects are detected, prompt notifications and relevant data are displayed, enabling the operator to respond quickly. The ultimate result is improved treatment safety, reduced complications, and enhanced patient experience and satisfaction.

[0014] In a second aspect, the present application provides a laser semiconductor array spectral matching optimization system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the laser semiconductor array spectral matching optimization system to perform the method described in the first aspect and any possible implementation method of the first aspect.

[0015] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a laser semiconductor array spectral matching optimization system, causes the laser semiconductor array spectral matching optimization system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0016] In a fourth aspect, the present application provides a computer program product, which, when running on a laser semiconductor array spectral matching optimization system, enables the laser semiconductor array spectral matching optimization system to perform the method described in the first aspect and any possible implementation of the first aspect.

[0017] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Due to the adoption of closed-loop control technology that integrates large models to analyze personalized medical aesthetic functional requirements and combines spectral detection with deep learning optimization models, it effectively solves the technical problem in existing technologies that the spectral output of laser semiconductor arrays is difficult to accurately match personalized medical aesthetic needs, thereby achieving the technical effect of greatly improving spectral matching accuracy, providing more accurate energy output for medical aesthetic treatment, and significantly improving treatment effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of a laser semiconductor array spectrum matching optimization method integrating an AI large model in an embodiment of the present application; Figure 2 This is another flow chart of the laser semiconductor array spectral matching optimization method integrating the AI large model in an embodiment of the present application; Figure 3 This is a schematic diagram of the structure of a physical device of the laser semiconductor array spectrum matching optimization system in the embodiment of the present application. DETAILED DESCRIPTION

[0019] The terms used in the following examples of the present application are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular expressions "a," "an," "said," "above," "the," and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to and encompasses any or all possible combinations of one or more of the listed items.

[0020] For ease of understanding, the following describes the process of the method provided by this implementation. Figure 1 , which is a flow chart of the laser semiconductor array spectral matching optimization method integrating the AI large model in the embodiment of the present application.

[0021] S101. Obtaining the user's personalized medical beauty function demand information; Among them, "users" refer to individuals who receive medical beauty treatments, and are people with different skin types, ages, beauty demands and other characteristics; "personalized medical beauty function demand information" refers to the unique requirements of each user for medical beauty treatment functions based on their own situation. For example, some people want to remove facial freckles, and some people want to improve eye wrinkles.

[0022] This step is executed when the system is ready to provide medical aesthetic services to users. Specifically, the system collects user information through various interactive methods. First, it provides a visual interactive interface designed with a guided questionnaire layout, presenting common medical aesthetic need categories in modules, such as skin improvement, hair treatment, anti-aging and wrinkle removal. Each category is further subdivided into specific items, and users complete their initial selection by checking boxes or filling in text boxes.

[0023] In some embodiments, a user's personalized medical aesthetics needs can be obtained through various methods: Optionally, the system utilizes voice interaction technology, activates voice recognition, converts user speech into text, and then performs semantic analysis to identify key needs. This information is then combined with the user's historical data (if available) and similar case data within the system to generate personalized medical aesthetics needs. Optionally, the system connects with a medical institution's information management system (with user authorization) to directly obtain skin problem diagnosis information and past treatment recommendations from the user's medical history, integrating this information with the user's current active input to determine personalized medical aesthetics needs. It is understood that other methods can also be used, such as analyzing a user's search history and browsing behavior on medical aesthetics platforms to determine needs, which are not limited here. In addition, during the information collection process, data security must be ensured, and encrypted transmission and storage technologies must be used to protect user privacy.

[0024] In some embodiments, before this step, a plurality of medical aesthetic function terms and corresponding spectral parameter information sets may be obtained. The spectral parameter information includes at least wavelength range, intensity threshold, and pulse frequency. Specifically, the system first collects authoritative lists of medical aesthetic function terms from medical literature, equipment manufacturer specifications, and clinical guidelines, standardizes them, and eliminates synonyms and ambiguous expressions. Subsequently, for each functional term, the corresponding spectral parameter information set is collected. These data come from a large number of clinical experiments and equipment parameter calibrations. The mapping relationship between the medical aesthetic function terms and the spectral parameter information is then reconstructed. Specifically, the system first performs feature engineering on the spectral parameter information, discretizing the continuous parameter values into feature vectors, such as dividing the wavelength range into several subintervals and mapping the intensity threshold to a logarithmic coordinate system. Then, for each medical aesthetic function term, the system uses a machine learning algorithm or expert rules to construct a mapping between it and the spectral parameter features. For deterministic mappings, the system uses regression analysis to establish functional relationships between functional terms and parameters. For probabilistic mappings, the system constructs conditional probability tables or Bayesian networks to calculate the probability of occurrence of various parameter combinations under a given function. Based on these mappings, a spectral medical aesthetics functional map is constructed. Specifically, the system first converts medical aesthetics functional terms and spectral parameter information into graph nodes, assigning each node a unique identifier and adding attribute descriptions. For example, attributes for a "laser freckle removal" node might include applicable skin types and common treatment areas; attributes for a wavelength parameter node might include physical units and safety ranges. Then, based on the established mapping relationships, directed edges are created between the nodes, with edge weights representing the strength or confidence of the mapping. The system also incorporates ontology concepts to define a type hierarchy for nodes and edges. For example, a "wavelength range" node could be classified as a "physical parameter" type, and a "laser freckle removal" node could be classified as a "treatment function" type. Once the map is constructed, inference rules are added to support graph-based query and reasoning. For example, the path search algorithm can be used to find the medical aesthetic function most relevant to specific spectral parameters, or semantic reasoning can be used to predict new function-parameter mapping relationships. In this way, for common medical aesthetic projects, the system can directly recommend parameters, simplify the operation process, and improve work efficiency.

[0025] In some embodiments, the user's skin image information can be first obtained, and skin feature data can be determined based on the skin image information. Specifically, this is usually performed before treatment, using the system's high-resolution camera and other image acquisition equipment to capture the user's skin from multiple angles under standard lighting conditions. During shooting, the camera automatically focuses and adjusts exposure parameters to ensure clear images and accurate color reproduction. The acquired raw image data is transmitted to the system for preprocessing. After acquiring the skin image information, the system first extracts features from the preprocessed skin image, applying computer vision algorithms to identify features such as the texture and color distribution of the skin surface. The extracted features are then analyzed and quantified using machine learning models, such as using a convolutional neural network (CNN) to detect wrinkles and pores and using color space analysis to assess skin tone evenness. Ultimately, the analysis results are converted into specific skin feature data, forming a digital description of the user's skin condition.

[0026] The system then determines desired function recommendations based on the skin feature data and a pre-set function recommendation model. This function recommendation model is pre-trained using machine learning on multiple sample sets annotated with skin feature data, actual medical aesthetic function selections, and feedback on medical aesthetic results. "Skin feature data" refers to a set of parameters that quantitatively describe skin condition, extracted through analysis of skin image information, including but not limited to skin type (e.g., oily, dry, combination), skin tone evenness, wrinkle depth, pore size, and pigmentation. The "function recommendation model" is an intelligent model built using a machine learning algorithm to predict the medical aesthetic functions a user may need based on the skin feature data. This model is pre-trained using multiple sample sets annotated with skin feature data, actual medical aesthetic function selections, and feedback on medical aesthetic results. The "desired function recommendations" are the model's output, personalized medical aesthetic function recommendations tailored to the user's skin condition. This step is performed after the skin feature data is determined. Specifically, the system inputs the acquired skin feature data into the pre-trained function recommendation model. The model then conducts an in-depth analysis of the input skin feature data based on the learned correlations between skin features and medical aesthetic functions. First, the model weights various skin feature data. Weights are determined based on clinical experience and statistical results from extensive training data. For example, for skin with significant pigmentation, pigment-related features are weighted more highly. The model then calculates a recommendation score for each medical aesthetic function using its internal neural network structure or other machine learning algorithms. These scores reflect the effectiveness and applicability of each function in improving the user's current skin condition.

[0027] Finally, the system determines the final personalized medical aesthetic function requirements based on the recommended functions and the user-entered function requirements. This step occurs after the recommended functions are generated. Specifically, the system first compares and analyzes the recommended functions with the user-entered function requirements, identifying any overlaps and differences. For overlapping function requirements, the system directly incorporates them into the final solution and refines them based on the priorities and parameter recommendations in the recommended functions. For any discrepancies, the system conducts a deeper analysis. If a user-entered function requirement is not included in the recommended functions, the system evaluates its plausibility and feasibility. For example, if a user requests a specific laser treatment, but the recommended functions indicate that it is unsuitable for their skin condition, the system will explain the recommendation and suggest alternative options through a user-computer interface. If the recommended functions include important functions that the user did not mention, the system will also explain these to the user, emphasizing their necessity for improving their overall skin condition. After thorough consultation with the user, the system will make adjustments based on their feedback. If the user insists on their original request and it is medically feasible, the system will incorporate it into the final solution, but will adjust the relevant parameters appropriately to ensure safety and effectiveness. Ultimately, the system integrates all information to generate the final personalized medical beauty function demand information. This information not only takes into account the user's skin characteristics and objective needs, but also fully respects the user's subjective wishes and preferences, achieving an organic combination of personalization and professionalism.

[0028] S102, determining spectrum requirement information based on the personalized medical aesthetic function requirement information; Specifically, the personalized medical aesthetic function demand information is input into a pre-built large language model to determine the medical aesthetic function terms and modifiers in the personalized medical aesthetic function demand information, and the modifiers include at least part qualifiers and degree qualifiers; combined with the medical aesthetic function terms, the basic spectral parameter information set corresponding to the medical aesthetic function terms is determined based on the spectral medical aesthetic function map; the basic spectral parameter information set is parameter-corrected according to the modifiers to generate spectral demand information.

[0029] Among them, "medical aesthetic function terms" refer to professional vocabulary that describes specific medical aesthetic functions, such as "freckle removal" and "skin rejuvenation"; "modifiers" are vocabulary used to limit or refine medical aesthetic function terms, among which "part qualifiers" are used to indicate specific body parts for medical aesthetic treatment, such as "face" and "eyes", and "degree qualifiers" are used to describe the degree of medical aesthetic effects, such as "mild" and "deep"; "spectral medical aesthetic function atlas" is a pre-built knowledge base that establishes a mapping relationship between medical aesthetic function terms and basic spectral parameter information; "basic spectral parameter information set" refers to a set of basic spectral parameters corresponding to specific medical aesthetic function terms, including wavelength range, intensity threshold, etc.; "parameter correction" refers to the process of adjusting and optimizing basic spectral parameter information according to modifiers; "spectral demand information" is the final set of spectral parameters that meets the user's personalized medical aesthetic function needs.

[0030] Specifically, the personalized medical aesthetic function demand information is pre-processed to remove typos, extra spaces, and other errors, and to unify the text format. After pre-processing, the information is input into the model. The multi-head attention mechanism within the model captures the contextual information of each word in the text and, through self-attention calculations, understands the relationship between each word and other words. For example, if the input is "remove deep dark circles under the eyes", the attention mechanism can analyze the semantic relationship between "remove", "dark circles", "eyes", and "deep". Then, the model's understanding of the text gradually deepens, and finally outputs the recognition result through the classification layer, determining that "remove dark circles" is a medical aesthetic function term, "eyes" is a location qualifier, and "deep" is a degree qualifier.

[0031] We selected a general pre-trained large language model and collected a large amount of text data from the medical aesthetics field, including descriptions of various medical aesthetics function requirements and professional literature, to construct a dataset specific to this field. This dataset was used to fine-tune the pre-trained model. During the fine-tuning process, the model learned the language patterns and terminology specific to this field. When fed personalized medical aesthetics function requirement information, the model can more accurately identify medical aesthetics function terms and modifiers. For example, for the phrase "improve dark corners of the mouth," the fine-tuned model accurately identifies "improve dark corners of the mouth" as a medical aesthetics function term and "corners of the mouth" as a location qualifier.

[0032] After acquiring medical aesthetic function terms, the system uses the query language of the graph database to query the corresponding spectral parameter information nodes, obtaining a basic spectral parameter information set, such as specific wavelength ranges and energy density. This basic spectral parameter information set is then modified based on the modifiers to generate spectral requirement information. The specific parameter modification process is as follows: If the modifier includes a site qualifier, the tissue structural parameters corresponding to the site qualifier, including epidermal thickness, dermal depth, and local pigment density, are extracted from a pre-set site optical properties database. Specifically, the system pre-builds a large and detailed site optical properties database, which stores tissue structural parameters such as epidermal thickness, dermal depth, and local pigment density for various human body parts. When receiving a modifier containing a site qualifier, the system uses keyword matching technology to quickly extract the corresponding parameter information from the database. For example, if the modifier is "face," the system uses an efficient search algorithm to accurately locate facial-related tissue structural parameter data.

[0033] The optical attenuation coefficient of the target site is calculated based on tissue structural parameters. Specifically, a specific optical attenuation calculation model is used. This model is typically constructed based on the theory of light propagation in biological tissue, taking into account the tissue's absorption and scattering properties. Based on the Beer-Lambert law, this model is modified based on the tissue structural parameters of the target site. For example, during the calculation, epidermal thickness, dermal depth, and local pigment density are substituted as variables into the formula: Optical attenuation coefficient = Absorption coefficient + Scattering coefficient. The absorption coefficient is related to the local pigment density, while the scattering coefficient is related to the tissue's microstructure (influenced by epidermal thickness and dermal depth). This calculation method can accurately determine the attenuation of a target site for a specific wavelength of light.

[0034] Based on the optical attenuation coefficient, the wavelength of the spectrum in the basic spectral parameter information set is adaptively adjusted for the specific site. Simultaneously, the spectral intensity in the basic spectral parameter information set is adjusted accordingly based on the biological safety threshold of the target site. Specifically, a wavelength adjustment algorithm is used to determine the appropriate wavelength adjustment based on the relationship between the optical attenuation coefficient and wavelength. Generally speaking, a large optical attenuation coefficient indicates rapid energy decay at that site, necessitating an appropriate increase in the wavelength to ensure laser penetration into the target tissue. Simultaneously, the spectral intensity is adjusted based on the biological safety threshold of the target site. This biological safety threshold has been determined through extensive experimental and clinical research, and the system pre-stores safety threshold data for each site. During intensity adjustment, an intensity control algorithm is employed. This algorithm compares the current spectral intensity with the safety threshold. If the intensity exceeds the threshold, it is attenuated according to a specific ratio. If the intensity is too low, it is increased appropriately, while always ensuring that it does not exceed the safety threshold. The adjusted spectral wavelength and intensity are then integrated to form the final spectral requirement information tailored to the target site. This process is achieved through data fusion technology, which combines wavelength and intensity data in a specific format to generate a complete spectral requirement data set, providing an accurate basis for the subsequent determination of the operating parameters of the laser semiconductor array.

[0035] If the modifier contains a degree qualifier, the degree qualifier is mapped to a predefined treatment depth level. Specifically, a mapping table of degree qualifiers and treatment depth levels can be established in advance, and the mapping table is developed based on clinical experience and medical research. When the system recognizes that a modifier contains a degree qualifier, the degree qualifier is converted into the corresponding treatment depth level by looking up the mapping table. For example, "mild" corresponds to a shallow treatment depth level, "moderate" corresponds to a medium treatment depth level, and "severe" corresponds to a deep treatment depth level. This mapping relationship provides a clear direction for subsequent parameter adjustments.

[0036] Based on the treatment depth level, the corresponding parameter adjustment rules are extracted from a pre-set depth adjustment strategy library. Specifically, the depth adjustment strategy library is a database that stores spectral parameter adjustment rules for different treatment depth levels. These rules are derived from the analysis and summary of a large number of clinical cases and experimental data. The system uses a rule retrieval algorithm to quickly find the corresponding adjustment rules based on the treatment depth level. For example, for deep treatment depth levels, the rules may dictate a reduction in pulse frequency and an increase in energy density.

[0037] According to the parameter adjustment rules, the pulse frequency in the basic spectral parameter information set is adjusted. Specifically, a pulse frequency adjustment algorithm is used to increase or decrease the pulse frequency according to the requirements of the rules. For example, if the rules require a reduction in the pulse frequency, the algorithm will reduce it according to a certain proportion based on the current pulse frequency value. Based on the corrected pulse frequency and target depth, the corresponding energy input is calculated. An energy calculation model is used here, which takes into account factors such as pulse frequency, target depth, and energy absorption characteristics of tissue. Based on the calculated energy input, an energy density adjustment algorithm is used to adjust the energy density in the basic spectral parameter information set to ensure that the energy density meets the treatment needs and is within a safe range.

[0038] Based on the corrected pulse frequency and target depth, the corresponding energy input is calculated and the energy density in the basic spectral parameter information set is adjusted. Specifically, a pre-established energy density adjustment reference table can be created in the system. This table contains energy density adjustment values corresponding to different pulse frequency and target depth combinations. The data for this reference table is derived from a large amount of clinical trial and actual treatment case data. During experiments and practice, energy density data that achieves the desired treatment effect at different pulse frequencies and target depths are recorded and compiled and analyzed to form a reference table. Once the corrected pulse frequency and target depth are obtained, the system queries the energy density adjustment reference table based on these two parameters. If a perfect match exists in the reference table for the pulse frequency and target depth combination, the corresponding energy density adjustment value is directly obtained. If not, an interpolation method is used to estimate the energy density. Once the energy density adjustment value is obtained, the system adjusts the energy density in the basic spectral parameter information set. If the calculated adjustment value increases the energy density, the system increases the set energy density by a certain percentage; if it decreases the energy density, the set energy density is reduced by the corresponding percentage. During the adjustment process, the system monitors energy density changes in real time to ensure that the adjusted energy density remains within a safe and effective range. Finally, the adjusted pulse frequency and energy density are combined with other relevant spectral parameters. These include wavelength and intensity, which were previously adjusted based on site qualifiers or other factors. The system collects these parameters in preparation for subsequent integration.

[0039] S103, determining operating parameter information of the laser semiconductor array using a preset laser semiconductor array matching strategy library in combination with the spectral requirement information, wherein the laser semiconductor array matching strategy library is constructed in advance based on multiple spectral requirement information sets annotated with the laser semiconductor array operating parameter information; The "pre-set laser semiconductor array matching strategy library" refers to a pre-established database that stores the correspondence between spectral requirements and laser semiconductor array operating parameters. These relationships are constructed based on a large set of spectral requirements information annotated with laser semiconductor array operating parameter information. The "laser semiconductor array operating parameter information" covers the various parameters that drive the laser semiconductor array, such as current, voltage, and temperature control parameters, which determine the output characteristics of the laser semiconductor array.

[0040] This step is performed after the spectral requirement information is determined, and is used to provide appropriate working parameter settings for the laser semiconductor array to meet the spectral requirements of medical aesthetic treatment. Specifically, after the system obtains the spectral requirement information, it will use it as a search condition to search in the preset laser semiconductor array matching strategy library. The strategy library stores a large amount of correspondence between spectral requirement information and laser semiconductor array working parameters. These relationships are obtained through the analysis and research of a large amount of historical data. The system uses an efficient search algorithm to quickly locate the working parameter information that matches the current spectral requirement information. For example, if the spectral requirement information requires a specific wavelength range and intensity threshold, the system will search the strategy library for all records that meet the wavelength and intensity requirements, and screen out the most suitable laser semiconductor array working parameter combination.

[0041] S104, obtaining output spectrum information of the laser semiconductor array in real time through a spectrum detection device; "Spectral detection equipment" refers to an instrument used to measure the spectral characteristics of light emitted by a semiconductor laser array. It can detect parameters such as wavelength distribution and intensity. "Output spectrum information" refers to parameter data related to the spectrum actually emitted by the semiconductor laser array, reflecting the actual laser output.

[0042] This step is continuously executed after the laser semiconductor array starts working. Its purpose is to monitor the output status of the laser in real time and provide a basis for subsequent parameter adjustments. Specifically, the spectrum detection device is connected to the output end of the laser semiconductor array. When the laser semiconductor array emits laser light, the spectrum detection device starts working. The optical sensor inside the device captures the laser light and converts it into an electrical signal. After amplification, filtering and other processing, these electrical signals are transmitted to the data processing unit. The data processing unit analyzes and calculates the electrical signals according to a pre-set algorithm to obtain spectral information such as the wavelength distribution and intensity of the laser. For example, by analyzing the frequency of the electrical signal, the wavelength of the laser can be determined; by measuring the signal intensity, the intensity value of the spectrum can be obtained. This information will be recorded and transmitted to the subsequent processing system in real time to ensure timely understanding of the laser output.

[0043] S105. Determine parameter adjustment information for the laser semiconductor array using a spectrum matching optimization model based on the spectrum requirement information and the output spectrum information, wherein the spectrum matching optimization model is constructed in advance through deep learning based on a plurality of spectrum requirement information and output spectrum information sets annotated with the laser semiconductor array parameter adjustment information; Among them, "output spectrum information" refers to the spectrum-related parameter data actually emitted by the laser semiconductor array obtained in real time by the spectrum detection equipment, which is used to reflect the actual output of the laser; "spectral matching optimization model" refers to a model constructed by deep learning based on multiple spectrum demand information and output spectrum information sets annotated with laser semiconductor array parameter adjustment information, which is used to analyze the difference between the spectrum demand information and the output spectrum information, and determine the parameter adjustment information to make the output spectrum more consistent with the demand spectrum; "parameter adjustment information" is used to represent the relevant information output by the spectrum matching optimization model for adjusting the working parameters of the laser semiconductor array, such as the adjustment amount of current, voltage, temperature control parameters, etc.

[0044] This step is performed after obtaining the required and output spectrum information. This occurs when the laser diode array is operating and the spectral detection equipment is monitoring the output spectrum in real time. The goal is to analyze the discrepancy between the actual output spectrum and the required spectrum and determine how to adjust the laser diode array's operating parameters to achieve a more accurate spectral match. Specifically, the system uses the acquired spectral requirement and output spectrum information as input to a spectral matching optimization model. During the model's construction phase, deep learning training is performed on a large number of sets of spectral requirement and output spectrum information annotated with parameter adjustment information. This model learns the optimal parameter adjustment methods for different discrepancies between spectral requirements and output spectra. During runtime, the model extracts and analyzes features from the input spectral requirement and output spectrum information, comparing the differences between the two in key parameters such as wavelength, intensity, and pulse frequency. Based on these differences, the model uses the trained algorithms and weights to calculate parameter adjustments that can narrow these discrepancies. For example, the model determines the magnitude and direction of adjustments to operating parameters such as current, voltage, and temperature, thereby bringing the subsequent output spectrum of the laser diode array closer to the required spectrum.

[0045] S106 , adjusting the operating parameters of the laser semiconductor array based on the parameter adjustment information.

[0046] This step is performed after determining the parameter adjustment information. This occurs when the spectral matching optimization model has calculated that the operating parameters of the laser semiconductor array need to be adjusted. The goal is to modify the operating state of the laser semiconductor array based on the adjustment information, so that the output spectrum better meets the user's personalized medical aesthetic requirements. Specifically, after receiving the parameter adjustment information output by the spectral matching optimization model, the system parses it. Depending on the different operating parameters mentioned in the adjustment information, the corresponding adjustment operation is performed. If the adjustment information requires an increase in current, the system gradually increases the current supplied to the laser semiconductor array by a specified amount through the driver circuit. If the voltage is adjusted, the voltage regulation module precisely changes the voltage applied to the laser semiconductor array. For temperature control parameters, the system controls temperature control devices, such as a cooler or heater, to maintain the operating temperature of the laser semiconductor array at the value specified in the adjustment information. During the adjustment process, the system monitors the operating state of the laser semiconductor array in real time to ensure that the adjustments to various operating parameters are within a safe range and do not cause damage to the equipment. Furthermore, after the adjustment operation is completed, the system again obtains the output spectrum information through the spectral detection equipment to verify that the adjustment has achieved the expected results.

[0047] In the embodiments of the present application, due to the use of a closed-loop control technology that integrates a large model to analyze personalized medical aesthetic functional needs, combines spectral detection with a deep learning optimization model, and fine-tunes the spectral parameters based on the tissue structure parameters of the site and the treatment depth level, and uses a dual feedback mechanism that combines real-time monitoring of biosensors with micro-expression feedback model analysis, it is possible to accurately locate user needs, ensure parameter accuracy, achieve personalized adjustment of spectral parameters, and promptly detect negative effects during treatment. This effectively solves the problems in the prior art that the spectral output of laser semiconductor arrays is difficult to accurately match personalized medical aesthetic needs, it is difficult to take into account the characteristics of different body parts and different treatment degree requirements, and it is difficult to promptly detect negative effects on patients during laser treatment. This greatly improves the accuracy of spectral matching, provides more accurate energy output for medical aesthetic treatment, significantly improves treatment effects, generates spectral demand information that is highly adapted to the target site and treatment degree, and improves light energy utilization.

[0048] In some embodiments, during laser treatment for users, due to the lack of an effective real-time monitoring and feedback mechanism, relying solely on preliminary communication and preset parameters before treatment, it is difficult to fully and accurately understand the user's true feelings and physical reactions during treatment. This makes it impossible to adjust the treatment plan in a timely manner based on the user's actual experience and physical condition, and thus it is difficult to provide users with more accurate and personalized services. In this case, the following solution can be used to improve the situation. The following is a more detailed description of the process of the method provided in this implementation. Please refer to Figure 2, which is another flow chart of the laser semiconductor array spectral matching optimization method integrating the AI large model in the embodiment of the present application.

[0049] S201, monitoring status information of a treatment site of a patient in real time through a biosensor, wherein the status information includes at least tissue temperature; Among them, "biosensor" refers to a device that can detect physical and chemical changes in biological substances or biological-related substances and convert them into output signals. It can sense changes in the physiological state of the patient's treatment area; "patient treatment area" refers to the specific body part undergoing laser medical beauty treatment, such as the face, neck, etc.; "status information" refers to data reflecting the real-time physiological condition of the patient's treatment area, which at least includes tissue temperature. "Tissue temperature" refers to the temperature value inside the tissue of the treatment area, and its changes can intuitively reflect the thermal effect of laser treatment on the tissue.

[0050] This step is performed continuously during the laser diode array's aesthetic treatment of the patient. This occurs while the laser is being applied to the patient's treatment area, requiring real-time monitoring of the physiological state of the treatment area. Specifically, before treatment begins, the operator will carefully position the biosensor near the patient's treatment area to ensure it can accurately sense physiological changes there. The biosensor integrates multiple sensing elements. For tissue temperature monitoring, these typically use temperature-sensitive components such as thermistors and thermocouples. When the laser is applied to the treatment area, the tissue temperature changes due to absorption of the laser energy. These temperature changes cause changes in the physical properties of the sensor's sensing element. For example, the resistance of a thermistor changes with temperature. The sensor converts this change in physical properties into an electrical signal, whose parameters, such as amplitude and frequency, are a function of temperature. This electrical signal is then transmitted to a data acquisition device via wires or a wireless transmission module. The data acquisition device periodically samples the electrical signal at a preset sampling frequency, converting the continuous analog signal into a discrete digital signal for subsequent processing and analysis. This collected data is stored in real time in the system's memory and also transmitted to the data analysis module. The data analysis module will perform real-time analysis on the collected temperature data to determine whether the tissue temperature is within a safe range. If it exceeds the safety threshold, the system will trigger the corresponding early warning mechanism to remind the operator to take measures to prevent tissue damage due to overheating.

[0051] S202, obtaining facial expression image data of the patient; Among them, "facial expression image data" refers to image information obtained through image acquisition equipment that reflects changes in the patient's facial expressions. These images contain various facial expression characteristics of the patient during the treatment process, such as frowning, closing eyes, changes in the corners of the mouth, etc., which can intuitively show the patient's emotions and feelings.

[0052] This step is also performed continuously during medical aesthetic treatments using a laser semiconductor array. This is especially important in situations where real-time understanding of the patient's emotions and feelings is crucial to adjust the treatment plan. Specifically, specialized image acquisition equipment, typically a high-definition camera, is installed in the treatment room, strategically positioned to ensure a clear view of the patient's face. Before treatment begins, the operator will debug the camera, adjusting parameters such as focal length, aperture, and shooting angle to ensure a clear and complete facial image, accurately capturing subtle changes in the patient's facial expressions. Once laser treatment begins, the camera captures images at a preset frame rate, typically 25 or 30 frames per second. This ensures that the captured image data consistently reflects the dynamic changes in the patient's facial expressions. The captured image data is stored as digital signals in the system's storage device, typically in common image formats such as JPEG and PNG. These formats ensure image quality while effectively compressing the data for easy storage and transmission. At the same time, the image data will be transmitted in real time to the subsequent image processing module, which will pre-process the image, including grayscale, noise reduction, image enhancement and other operations to improve the clarity and recognizability of the image, and prepare for the subsequent analysis of the patient's adverse feedback data through the micro-expression feedback model.

[0053] S203, combining the facial expression image data with the patient's adverse feedback data using a micro-expression feedback model, wherein the micro-expression feedback model is constructed by deep learning based on a plurality of micro-expression image sets annotated with corresponding adverse feedback data under different treatment conditions; Among them, "facial expression image data" refers to the image information obtained in step S202 that reflects the changes in the patient's facial expression; "micro-expression feedback model" refers to a model constructed through deep learning, which is used to analyze facial expression image data, identify micro-expression features therein, and determine the patient's negative feedback data based on these features; "adverse feedback data" is used to represent relevant data on negative feelings such as discomfort and pain that may occur in patients during laser treatment, such as pain level, discomfort type, etc.; "micro-expression image set" refers to a large collection of facial images containing various micro-expressions, which are annotated with corresponding negative feedback data and serve as basic data for training the micro-expression feedback model.

[0054] This step is performed after acquiring the patient's facial expression image data. During ongoing laser treatment, the system needs to analyze the patient's facial expressions to determine whether they exhibit any negative feelings. Specifically, after acquiring the facial expression image data, the system inputs this data into the micro-expression feedback model. To construct the micro-expression feedback model, a large amount of micro-expression image data from various treatment scenarios was collected. These images were annotated by professionals, with corresponding negative feedback data, such as mild pain and moderate discomfort. This annotated data was then trained using a deep learning algorithm, such as a convolutional neural network (CNN). During training, the CNN model automatically learns the relationship between features in the micro-expression images, such as subtle facial muscle movements and eye changes, and negative feedback data. When new facial expression image data is input, the model performs feature extraction, converting the image's pixel information into an abstract feature vector. The model then uses a classifier trained within the model to classify the feature vector and determine the type of negative feedback data corresponding to the micro-expression. For example, the model may determine that the patient's expression indicates that he or she is feeling mild pain. The system will then record the corresponding negative feedback data and combine it with the status information obtained from the biosensor (such as tissue temperature) to comprehensively determine whether the current laser treatment has a negative impact on the patient.

[0055] S204, combining the state information and the adverse feedback data to determine whether the current laser treatment has a negative impact on the patient; "Laser therapy" refers to the process of using a specific spectrum emitted by a laser semiconductor array to perform medical aesthetic treatments on patients; "negative effects" is used to refer to the adverse effects of laser therapy on the patient's body, including but not limited to tissue damage, excessive pain, allergic skin reactions, etc.

[0056] This step is performed after status information and negative feedback data are acquired. During ongoing laser treatment, the system needs to promptly determine whether the treatment has had any negative effects on the patient so that appropriate adjustments can be made. Specifically, the system first integrates and processes the status information and negative feedback data. For status information, it analyzes tissue temperature trends. If tissue temperature rises rapidly within a short period of time and exceeds the threshold corresponding to the normal physiological range, this may indicate excessive laser energy deposition in the local tissue, causing excessive thermal effects and increasing the risk of tissue damage. Negative feedback data is evaluated based on the pain severity and type of discomfort provided. If the patient reports moderate or higher pain levels, or experiences intense discomfort such as a burning sensation, these indicate a significant negative experience during treatment. The system then integrates this information and uses pre-defined judgment rules to determine whether a negative effect has occurred. These judgment rules can be based on extensive clinical data and medical research. For example, if tissue temperature exceeds a safety threshold and the patient reports moderate or higher pain, a negative effect is considered. Alternatively, if the negative feedback data indicates discomfort related to skin allergies (such as itching, redness, and swelling), even if the tissue temperature does not exceed the threshold, a negative effect is also considered.

[0057] If a negative impact occurs, execute step S205; If no negative impact is generated, step S206 is executed.

[0058] S205: Sending an adverse impact prompt to the user terminal, and displaying the status information and adverse feedback data to the visual terminal; Among them, "user end" refers to the user operation terminal that interacts with the laser semiconductor array spectral matching optimization system, usually computers, tablets and other devices used by medical staff, which are used to receive various types of information sent by the system and perform corresponding operations; "adverse effect prompt" means that the system sends a reminder message to the user end after determining that laser treatment has a negative impact on the patient. It can be in the form of pop-up notifications, sound alarms, etc., with the purpose of informing medical staff of abnormal situations during the treatment process; "visual end" refers to a device that can present information in a visual manner.

[0059] This step is performed after the laser treatment has been determined to have a negative impact on the patient. This occurs in scenarios where the system has determined an anomaly during the treatment process, requiring prompt notification to medical staff and the provision of relevant data for decision-making. Specifically, once the system concludes a negative impact, it first sends a notification to the user using a pre-defined notification method. If the notification method is a pop-up notification, a prominent window will appear on the user's interface, displaying a message such as "Laser treatment has had a negative impact. Please review relevant data." If the notification method is an audio alert, a specific warning tone will play to attract the attention of medical staff. Simultaneously, the system organizes and formats status information and negative feedback data. Status information includes information such as tissue temperature and other physiological parameters, along with their trends, in clear and understandable tables or charts, such as a line chart showing tissue temperature changes over time. Negative feedback data is categorized and listed based on pain severity and discomfort type. The processed status information and negative feedback data are then sent to the visual terminal for display. On the visual end device, medical staff can intuitively see these data. By observing abnormal tissue temperature and patient discomfort feedback, they can quickly understand the problems that arise during the treatment process and provide a basis for subsequent adjustments to the treatment plan.

[0060] In some embodiments, spectrum adjustment requirement information input by the user can also be obtained; parameter adjustment data for the laser semiconductor array is determined based on the spectrum adjustment requirement information; and the operating parameters of the laser semiconductor array are adjusted based on the parameter adjustment data. This step is performed after an adverse effect prompt is issued to the user and status information and adverse feedback data are displayed. The scenario is that during laser treatment, the system detects possible negative effects, and after providing feedback to the user, it waits for the user to decide whether to adjust the treatment plan based on the actual situation. Specifically, after receiving the prompt and relevant data issued by the system, the user will enter the spectrum adjustment requirement information on the user-side device based on their own professional judgment, clinical experience, and the patient's real-time condition.

[0061] S206: Control the laser semiconductor array to continue operating according to the current operating parameters.

[0062] This step is executed after prompts are sent to the user end and data is displayed to the visual end. The scenario is that after checking the status information and adverse feedback data, medical staff evaluate that although there are certain abnormalities in the current laser treatment, they are within an acceptable range and decide to let the laser semiconductor array maintain its current working state to continue treatment.

[0063] The solution in the embodiments of the present application effectively solves the problems in the prior art of difficulty in timely detecting the negative impact on patients during laser treatment, and the inability to flexibly adjust the treatment plan according to the patient's real-time status, thereby achieving the technical effect of improving treatment safety, reducing complications, enhancing patient treatment experience and satisfaction, and optimizing the treatment process while ensuring the treatment effect.

[0064] The following describes the laser semiconductor array spectrum matching optimization system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of the physical device structure of the laser semiconductor array spectral matching optimization system in an embodiment of the present application.

[0065] It should be noted that Figure 3 The structure of the laser semiconductor array spectrum matching optimization system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. Figure 3 As shown, the laser semiconductor array spectral matching optimization system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes, such as the methods described in the above embodiments, based on programs stored in a read-only memory (ROM) 302 or programs loaded from a storage unit 308 into a random access memory (RAM) 303. RAM 303 also stores various programs and data required for system operation. CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0066] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, pushbutton switches, etc.; an output section 307 including a liquid crystal display (LCD), an audio output device, indicator lights, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 310 as needed, so that computer programs read from the media can be installed into the storage section 308 as needed. In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program stored on a computer-readable medium, the computer program containing the computer program for executing the method illustrated in the flowchart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication section 309 and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0067] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. Specifically, the laser semiconductor array spectral matching optimization system of this embodiment includes a processor and a memory, and a computer program is stored on the memory. When the computer program is executed by the processor, the laser semiconductor array spectral matching optimization method integrating the AI large model provided in the above embodiment is implemented.

[0068] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the laser semiconductor array spectral matching optimization system described in the above embodiments, or may exist independently and not be incorporated into the laser semiconductor array spectral matching optimization system. The storage medium carries one or more computer programs. When executed by a processor of the laser semiconductor array spectral matching optimization system, the laser semiconductor array spectral matching optimization system implements the laser semiconductor array spectral matching optimization method integrating the AI large model provided in the above embodiments.

Claims

1. A laser semiconductor array spectral matching optimization method integrating an AI large model, applied to a laser semiconductor array spectral matching optimization system, characterized in that: The method comprises: Obtain users' personalized medical beauty function demand information; Determining spectral demand information in combination with the personalized medical aesthetic function demand information; In combination with the spectral requirement information, the operating parameter information of the laser semiconductor array is determined by a preset laser semiconductor array matching strategy library, wherein the laser semiconductor array matching strategy library is constructed in advance based on multiple spectral requirement information sets annotated with the operating parameter information of the laser semiconductor array; The output spectrum information of the laser semiconductor array is obtained in real time through a spectrum detection device; Combining the spectral requirement information and the output spectral information, determining parameter adjustment information of the laser semiconductor array through a spectral matching optimization model, wherein the spectral matching optimization model is constructed through deep learning based on multiple sets of spectral requirement information and output spectral information annotated with the parameter adjustment information of the laser semiconductor array; The operating parameters of the laser semiconductor array are adjusted in combination with the parameter adjustment information.

2. The method according to claim 1, characterized in that Before obtaining the user's personalized medical beauty function demand information, it also includes: Acquire a plurality of medical aesthetic function terms and corresponding spectral parameter information sets, wherein the spectral parameter information includes at least a wavelength range, an intensity threshold, and a pulse frequency; Constructing a mapping relationship between the medical aesthetic function term and the spectral parameter information; Based on the mapping relationship, a spectral medical aesthetic function map is constructed.

3. The method according to claim 1, characterized in that The step of determining the spectrum requirement information in combination with the personalized medical aesthetic function requirement information specifically includes: Inputting the personalized medical aesthetic function requirement information into a pre-built large language model to determine medical aesthetic function terms and modifiers in the personalized medical aesthetic function requirement information, wherein the modifiers include at least part qualifiers and degree qualifiers; In combination with the medical aesthetic function term, determining a basic spectral parameter information set corresponding to the medical aesthetic function term based on a spectral medical aesthetic function map; Parameter modification is performed on the basic spectrum parameter information set according to the modifier to generate spectrum requirement information.

4. The method according to claim 3, characterized in that The step of modifying the basic spectrum parameter information set according to the modifier to generate spectrum requirement information specifically includes: If the modifier includes a site qualifier, then extracting tissue structure parameters corresponding to the site qualifier from a preset site optical property database, including epidermal thickness, dermal depth, and local pigment density; calculating the optical attenuation coefficient of the target site based on the tissue structure parameters; Adaptively adjusting the wavelength of the spectrum in the basic spectral parameter information set according to the optical attenuation coefficient, and adjusting the intensity of the spectrum in the basic spectral parameter information set according to the biological safety threshold of the target site; Integrate the adjusted spectral wavelength and spectral intensity to generate spectral requirement information suitable for the target area; If the modifier includes a degree qualifier, mapping the degree qualifier to a predefined treatment depth level; According to the treatment depth level, extracting corresponding parameter adjustment rules from a preset depth adjustment strategy library; adjusting the pulse frequency in the basic spectrum parameter information set according to the parameter adjustment rule; Calculating the corresponding energy input based on the corrected pulse frequency and target depth, and adjusting the energy density in the basic spectral parameter information set; The adjusted pulse frequency and energy density are integrated to generate spectral demand information adapted to the corresponding treatment level.

5. The method according to claim 1, wherein After the step of adjusting the operating parameters of the laser semiconductor array in combination with the parameter adjustment information, the method further includes: Real-time monitoring of status information of the patient's treatment site by a biosensor, wherein the status information includes at least tissue temperature; Acquiring facial expression image data of the patient; Determining the patient's adverse feedback data using a micro-expression feedback model based on the facial expression image data, wherein the micro-expression feedback model is constructed through deep learning based on a plurality of micro-expression image sets previously labeled with corresponding adverse feedback data under different treatment conditions; Determining whether the current laser treatment has a negative impact on the patient by combining the status information and the adverse feedback data; If a negative impact occurs, an adverse impact prompt is issued to the user end, and the status information and adverse feedback data are displayed to the visual end.

6. The method according to claim 5, characterized in that If a negative impact is generated, after the steps of issuing a negative impact prompt to the user terminal and displaying the status information and negative feedback data to the visual terminal, the method further includes: Obtaining spectrum adjustment requirement information input by the user; Determining parameter adjustment data of the laser semiconductor array in combination with the spectrum adjustment requirement information; The operating parameters of the laser semiconductor array are adjusted in combination with the parameter adjustment data.

7. The method according to claim 1, characterized in that The steps for obtaining the user's personalized medical beauty function demand information include: Get the user's skin image information; determining skin feature data according to the skin image information; Determining the required function recommendations based on the skin feature data and a preset function recommendation model, wherein the function recommendation model is obtained by machine learning training based on multiple sample sets annotated with skin feature data, actual medical aesthetic function selection annotations, and medical aesthetic effect feedback annotations; The final personalized medical aesthetic function demand information is determined based on the demand function recommendation and the function demand input by the user.

8. A laser semiconductor array spectrum matching optimization system, characterized in that: The laser semiconductor array spectral matching optimization system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the laser semiconductor array spectral matching optimization system to perform the method described in any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a laser semiconductor array spectrum matching optimization system, the laser semiconductor array spectrum matching optimization system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on a laser semiconductor array spectrum matching optimization system, the laser semiconductor array spectrum matching optimization system is enabled to perform the method according to any one of claims 1 to 7.

Images