Fat melting method and system for automatically identifying skin features and storage medium

Abstract

The invention discloses a fat melting method and system for automatically recognizing skin features and a storage medium, and the method comprises the steps: collecting skin feature data, and constructing a three-dimensional surface model of a body surface region; inputting the skin feature data and the three-dimensional surface model into a pre-trained feature recognition model, and performing security level classification on the skin area by the feature recognition model to obtain a body surface classification three-dimensional model comprising a normal illumination area, an illumination limiting area and an illumination forbidding area; based on the depth gradient of the body surface grading three-dimensional model and the feature recognition confidence coefficient of the feature recognition model, dividing a first safety buffer zone for the normal illumination area and the illumination forbidding area, and dividing a second safety buffer zone for the illumination limiting area and the illumination forbidding area; generating an irradiation scheme according to the body surface grading three-dimensional model; and executing the irradiation operation according to the irradiation scheme.

Description

Technical Field

[0001] This invention relates to the field of non-invasive human tissue heating / fat reduction technology, belonging to the optical and control technology of medical / home beauty devices, and more specifically, to a fat melting method, system and storage medium that automatically identifies skin characteristics. Background Technology

[0002] As non-invasive liposuction technology becomes increasingly widely used in the beauty and health fields, users' demands for its automated operation and treatment safety continue to rise. However, existing liposuction solutions largely rely on the operator's experience to judge and exclude sensitive areas of the skin. This is not only highly subjective and lacks repeatability among different operators, but also fails to accurately distinguish between normal and sensitive areas. This can easily lead to the laser mistakenly targeting sensitive areas, resulting in local overheating damage. Furthermore, although existing liposuction technologies perform liposuction in sections for normal and sensitive areas, most do not set up a safe transition zone between the two. As a result, when the laser passes from a normal area to a sensitive area, the energy cannot be smoothly transferred, which can cause burns to the edges of sensitive areas or result in insufficient liposuction at the edges of normal areas.

[0003] Therefore, designing a fat reduction method, system, and storage medium that can automatically identify skin features, accurately divide safe-level areas, and coordinate and adjust the irradiation scheme has become an urgent problem to be solved in this field. Summary of the Invention

[0004] This invention is mainly achieved through the following technical solutions: This invention provides a method for automatically identifying skin features during fat reduction, comprising the following steps: Collect skin feature data and construct a three-dimensional surface model of the body surface area; Skin feature data and three-dimensional surface model are input into the pre-trained feature recognition model. The feature recognition model classifies the skin region according to the safety level and obtains a three-dimensional model of the body surface graded into normal irradiation area, limited irradiation area and prohibited irradiation area. Based on the depth gradient of the three-dimensional model of body surface grading and the feature recognition confidence of the feature recognition model, a first safety buffer zone is divided into normal irradiation zone and no-irradiation zone, and a second safety buffer zone is divided into limited irradiation zone and no-irradiation zone. Based on the three-dimensional model of body surface grading, an irradiation plan is generated; According to the irradiation plan, irradiation operations are carried out. The irradiation plan includes: irradiating the normally lit area according to the first mode, irradiating the restricted area according to the second mode; irradiating the first safety buffer zone according to the third mode, irradiating the second safety buffer zone according to the fourth mode, and shielding the prohibited area for irradiation. The energy output per unit time decreases progressively in the first, second, third, and fourth modes.

[0005] Furthermore, skin feature data includes one or more of the following: skin appearance data, deep skin structure data, skin temperature data, and three-dimensional morphology data of the skin surface.

[0006] Furthermore, the steps to obtain the three-dimensional model of the body surface grading include: Input skin feature data and 3D surface model into the feature recognition model; The feature recognition model marks the areas corresponding to skin appearance data that meet the preset high-risk standards as no-photography zones; The feature recognition model marks the areas corresponding to the deep skin structure data that meet the preset high-risk criteria as forbidden areas for exposure. The feature recognition model identifies areas corresponding to skin temperature data that meet preset high-risk standards as no-sunlight zones. The feature recognition model identifies areas corresponding to the three-dimensional morphological data of skin surfaces that meet the preset high-risk standards as no-photography zones.

[0007] Furthermore, the steps to obtain the three-dimensional model of the body surface grading also include: The feature recognition model identifies areas corresponding to skin appearance data that meet the preset medium-risk standards as restricted-light zones. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset medium-risk criteria as restricted exposure areas. The feature recognition model identifies areas corresponding to skin temperature data that meet preset medium-risk standards as restricted-sunlight zones. The feature recognition model identifies the areas corresponding to the three-dimensional morphological data of the skin surface that meet the preset medium-risk standards as restricted exposure areas.

[0008] Furthermore, the steps to obtain the three-dimensional model of the body surface grading also include: The feature recognition model identifies the areas corresponding to skin appearance data that meet the preset low-risk criteria as normal-light areas. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model identifies the areas corresponding to skin temperature data that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model identifies the areas corresponding to the three-dimensional morphology data of the skin surface that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model maps the normal illumination area, limited illumination area, and no-illumination area to the corresponding regions of the three-dimensional surface model, respectively, to obtain a graded three-dimensional model of the body surface.

[0009] Furthermore, the steps for defining the first safety buffer zone include: Obtain the boundary contours of the normal irradiation zone and the prohibited irradiation zone on the three-dimensional model of the body surface grading; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with the preset curvature threshold range, and the feature recognition confidence is matched and compared with the preset feature recognition threshold range. Adjust the width of the boundary contour based on the matching comparison results of confidence scores for surface curvature and / or feature identification; The adjusted boundary contour is used as the first safety buffer zone and mapped onto the corresponding area of ​​the body surface graded 3D model.

[0010] Furthermore, the steps for defining the second safety buffer zone include: Obtain the boundary contours of the restricted and prohibited light zones on the three-dimensional model of the body surface classification; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with the preset curvature threshold range, and the feature recognition confidence is matched and compared with the preset feature recognition threshold range. Adjust the width of the boundary contour based on the matching comparison results of confidence scores for surface curvature and / or feature identification; The adjusted boundary contour is used as a second safety buffer zone and mapped onto the corresponding area of ​​the body surface graded 3D model.

[0011] Furthermore, the steps for performing irradiation using a fat-dissolving laser also include: The three-dimensional model of body surface grading is divided into several sub-regions; Select non-adjacent sub-regions at preset intervals; Irradiate according to the pattern corresponding to the sub-region; After irradiation, the sub-region is cooled down.

[0012] Furthermore, the feature is that during the irradiation operation using the grease-melting laser, the grease-melting safety index data is monitored in real time. If an abnormality is detected in the grease-melting safety index data, the irradiation parameters of the corresponding area are adjusted or the irradiation is immediately suspended and a warning signal is triggered. The safety indicators for fat melting include one or more of the following: skin surface temperature, the fit between the device and the skin surface, feature recognition status, and laser power fluctuation.

[0013] Furthermore, real-time monitoring of grease melting safety indicators includes: The skin surface temperature is monitored in real time by a thermal infrared camera. If the skin surface temperature exceeds the preset temperature safety warning value, the irradiation will be stopped immediately and an audio-visual warning will be triggered.

[0014] Furthermore, real-time monitoring of grease melting safety indicators also includes: The device monitors the fit between the device and the skin surface in real time using a pressure sensor. If the fit is lower than the preset fit threshold, the irradiation will be paused immediately and an audio-visual alert will be triggered.

[0015] Furthermore, real-time monitoring of grease melting safety indicators also includes: The safety level identification status of each area is verified by a feature recognition model. If an abnormality in the safety level identification occurs, the irradiation parameters of the corresponding area are reduced, or irradiation is immediately suspended and a warning signal is triggered.

[0016] Furthermore, real-time monitoring of grease melting safety indicators also includes: The power fluctuation of the laser is monitored in real time by a power sensor. If the power fluctuation exceeds the preset power fluctuation range, the laser output will be stopped immediately and a prompt signal will be triggered.

[0017] Furthermore, the degreasing method also includes storing the irradiation parameters, degreasing safety index data, and abnormal handling information for each area in a log table after the irradiation operation is completed.

[0018] Furthermore, fat-dissolving methods also include: Extract the regions corresponding to security level anomalies; The regions are manually labeled with security levels and then added to the training set of the feature recognition model.

[0019] A fat-dissolving system that automatically identifies skin features includes: The data acquisition unit is configured to collect skin feature data of the area to be melted. The model building unit is configured to build a three-dimensional surface model of the body surface region; The safety level identification unit is configured to obtain a three-dimensional model of the body surface with three safety levels, including normal irradiation zone, limited irradiation zone and prohibited irradiation zone, based on skin feature data and three-dimensional surface model, and to divide a first safety buffer zone between the normal irradiation zone and the prohibited irradiation zone, and a second safety buffer zone between the limited irradiation zone and the prohibited irradiation zone. The laser control unit is configured to irradiate the normal illumination area, the limited illumination area, the first safety buffer zone, and the second safety buffer zone using the first to fourth modes respectively, and to shield the forbidden illumination area from irradiation. The energy output of the first to fourth modes decreases step by step within a unit time. The laser emitting unit is configured to irradiate the melting area according to the laser irradiation parameters; The cooling unit is configured to cool the melting area during the melting process.

[0020] Furthermore, the fat-dissolving system also includes: The safety indicator monitoring unit is configured to monitor the grease melting safety indicator data in real time during the irradiation operation. If abnormalities are detected in the grease melting safety indicator data, the laser irradiation parameters of the corresponding area will be adjusted or the irradiation will be immediately suspended and a warning signal will be triggered.

[0021] A storage medium that automatically identifies skin features, the storage medium storing instructions which, when invoked by a processor, are used to execute any of the above-mentioned fat-dissolving methods.

[0022] In summary, the present invention has the following advantages compared with the prior art: This solution collects multi-dimensional skin feature data and combines it with a pre-trained feature recognition model to achieve automated identification of local skin features. It accurately delineates prohibited, restricted, and normal irradiation areas and sets adaptive safety buffer zones. Compared with traditional manual identification, it completely eliminates the reliance on operator experience and effectively avoids the risk of missed detection of sensitive areas due to subjective misjudgment. It can achieve precise differentiated irradiation for areas with different safety levels. It avoids the problems of irradiation range deviation and intensity loss control caused by the lack of clear execution standards in traditional solutions. While preventing damage to sensitive areas, it can also ensure the fat reduction effect in other areas, further enhancing the safety and treatment targeting of the fat reduction process and adapting to the usage needs of different skin feature scenarios. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1 A flowchart illustrating the steps of the fat-dissolving method for automatically identifying skin features provided in this application. Figure 2 The flowchart of the steps for dividing the first safety buffer zone is provided in this application document; Figure 3 A structural diagram of the system for automatically recognizing skin features provided in this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0025] Firstly, such as Figure 1As shown, the present invention provides an automatic fat-dissolving method for identifying skin features, comprising the following steps: Collect skin feature data and construct a three-dimensional surface model of the body surface area; Skin feature data and three-dimensional surface model are input into the pre-trained feature recognition model. The feature recognition model classifies the skin region according to the safety level and obtains a three-dimensional model of the body surface graded into normal irradiation area, limited irradiation area and prohibited irradiation area. Based on the depth gradient of the three-dimensional model of body surface grading and the feature recognition confidence of the feature recognition model, a first safety buffer zone is divided into normal irradiation zone and no-irradiation zone, and a second safety buffer zone is divided into limited irradiation zone and no-irradiation zone. Based on the three-dimensional model of body surface grading, an irradiation plan is generated; According to the irradiation plan, irradiation operations are carried out. The irradiation plan includes: irradiating the normally lit area according to the first mode, irradiating the restricted area according to the second mode; irradiating the first safety buffer zone according to the third mode, irradiating the second safety buffer zone according to the fourth mode, and shielding the prohibited area for irradiation. The energy output per unit time decreases progressively in the first, second, third, and fourth modes.

[0026] In this embodiment, skin feature data is acquired through multimodal acquisition, specifically including: acquiring skin appearance data using a visible light camera with a resolution of 12 megapixels; acquiring blood vessel distribution characteristics in the deep skin structure data using a near-infrared camera with a working wavelength of 850~940nm; acquiring skin temperature data using a thermal infrared camera with a resolution of 320×240; and acquiring three-dimensional morphological data of the skin surface using a ToF depth camera or millimeter-wave radar. After acquisition, the three-dimensional morphological data of the skin surface is processed by point cloud filtering and time registration, and then input into the Poisson reconstruction algorithm to construct a three-dimensional surface model.

[0027] In this embodiment, based on the regional division results of the three-dimensional model of body surface grading, the patterns corresponding to different regions are determined. The irradiation parameters of the patterns include mask transmittance, laser output power, laser irradiation duty cycle, and laser action duration.

[0028] In the first mode, the preset range of mask transmittance is 80%~100%, the preset range of laser output power is 12~16W, the preset range of laser duty cycle is 60%~80%, and the preset range of single irradiation duration is 15~22s.

[0029] The second mode has a mask transmittance preset range of 60%~80%, a laser output power preset range of 8~11W, a laser duty cycle preset range of 40%~60%, and a single irradiation duration preset range of 18~28s.

[0030] The preset range for the mask transmittance in the third mode is 40%~60%, the preset range for laser output power is 5~8W, the preset range for laser duty cycle is 25%~45%, and the preset range for single irradiation duration is 20~30s.

[0031] The fourth mode has a preset range of mask transmittance of 20% to 40%, a preset range of laser output power of 3 to 6W, a preset range of laser duty cycle of 15% to 35%, and a preset range of single irradiation duration of 22 to 32 seconds. Under the above parameter configuration, the energy output per unit time corresponding to each mode decreases gradually as the safety level increases, thereby forming a gradual transition of energy from high to low near the no-light zone.

[0032] Then, the first, second, third, and fourth modes are used to irradiate the normal illumination area, the limited illumination area, the first safety buffer zone, and the second safety buffer zone, respectively. It should be noted that the specific ranges of mask transmittance, power, and single irradiation duration described above are merely preferred embodiments of the present invention under specific experimental conditions and are not intended to limit the scope of protection of the present invention. In practical applications, those skilled in the art can adjust the above parameters according to laser wavelength, tissue thermal characteristics, and individual differences, without departing from the logical framework of the present invention.

[0033] In one possible implementation, skin feature data includes one or more of the following: skin appearance data, deep skin structure data, skin temperature data, and three-dimensional morphology data of the skin surface.

[0034] In one possible implementation, the steps to obtain a three-dimensional model of the body surface grading include: Input skin feature data and 3D surface model into the feature recognition model; The feature recognition model marks the areas corresponding to skin appearance data that meet the preset high-risk standards as no-photography zones; The feature recognition model marks the areas corresponding to the deep skin structure data that meet the preset high-risk criteria as forbidden areas for exposure. The feature recognition model identifies areas corresponding to skin temperature data that meet preset high-risk standards as no-sunlight zones. The feature recognition model identifies areas corresponding to the three-dimensional morphological data of skin surfaces that meet the preset high-risk standards as no-photography zones.

[0035] In this embodiment, the feature recognition model calculates the melanin-related index (MI) of the skin appearance data. The calculation formula can be MI = 0.5 × (RB) / G + 0.3 × NIR / R, where R represents the red light intensity value, RB represents the blue light intensity value, and G represents the green light intensity value. Alternatively, a conventional calculation formula can be used. Where R represents the red light intensity value; the MI is compared with the preset melanin threshold. When the MI≥0.6, the corresponding area is marked as a forbidden area. This area generally represents sensitive areas such as suspected pigmented lesions, tattoo areas, or dark dye areas.

[0036] The feature recognition model extracts vascular distribution features from deep skin structure data, identifying vessels with a diameter ≥2mm and a density ≥3 vessels / mm. 2 When this happens, the corresponding area will be marked as a no-photography zone. This area generally represents sensitive areas such as varicose veins or vascular clusters.

[0037] The feature recognition model extracts skin temperature data. When the skin temperature data is ≥40℃, the corresponding area is marked as a no-light zone. This area generally represents an area of ​​inflammation, local infection, or abnormal heat stress.

[0038] The feature recognition model extracts the depth value (H) and equivalent diameter (D) from the three-dimensional morphology data of the skin surface. When H > 3 mm and D > 5 mm, the corresponding area is marked as a no-light zone. Such areas are generally the umbilicus or deep depressions. The above threshold division based on blood vessel diameter and density, depth value (H) and equivalent diameter (D) is a typical configuration of this embodiment, used to illustrate the principle of classifying different risk levels. For other cases that do not completely fall into the above interval combination, the feature recognition model makes a judgment based on the risk criteria obtained during training. It is usually based on the risk correlation of each skin feature data to comprehensively determine whether to classify it as a no-light zone according to the level of risk. This invention is not limited to the specific thresholds and interval divisions mentioned above.

[0039] In one possible implementation, the steps for obtaining the three-dimensional model of the body surface grading also include: The feature recognition model identifies areas corresponding to skin appearance data that meet the preset medium-risk standards as restricted-light zones. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset medium-risk criteria as restricted exposure areas. The feature recognition model identifies areas corresponding to skin temperature data that meet preset medium-risk standards as restricted-sunlight zones. The feature recognition model identifies the areas corresponding to the three-dimensional morphological data of the skin surface that meet the preset medium-risk standards as restricted exposure areas.

[0040] In this embodiment, when 0.4≤MI<0.6, the corresponding area is marked as a restricted area. Such areas are mostly mature scars, stretch marks, and other areas with slightly abnormal pigmentation but no high risk. When 1mm ≤ vessel diameter < 2mm and 1 vessel / mm 2 ≤Vascular density <3 vessels / mm 2 At that time, the corresponding area was marked as the restricted area, which is mostly an area with mild vascular exposure; When 37℃≤skin temperature<40℃, the corresponding area is marked as the restricted area. This area generally corresponds to the area of ​​mild congestion. When 1mm≤H≤3mm and 3mm≤D≤5mm, the corresponding area is marked as the restricted area. This area generally corresponds to the edge of skin tags or bony prominences.

[0041] In this embodiment, the feature recognition model is trained based on a large number of labeled samples, covering phantoms of different ages, genders and regions. The feature recognition model includes, but is not limited to, DNN deep neural network model, CNN convolutional neural network model, ResNet residual network model or Transformer architecture model. AI models that use one or more of the above models in combination to achieve skin safety level classification are all within the protection scope of this application.

[0042] In one possible implementation, the steps for obtaining the three-dimensional model of the body surface grading also include: The feature recognition model identifies the areas corresponding to skin appearance data that meet the preset low-risk criteria as normal-light areas. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model identifies the areas corresponding to skin temperature data that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model identifies the areas corresponding to the three-dimensional morphology data of the skin surface that meet the preset low-risk criteria as normal irradiation areas. The feature recognition model maps the normal illumination area, limited illumination area, and no-illumination area to the corresponding regions of the three-dimensional surface model, respectively, to obtain a graded three-dimensional model of the body surface.

[0043] In this embodiment, the area corresponding to MI < 0.4 is designated as the normal illumination area, and the area with a vessel diameter < 1 mm and a vessel density < 1 vessel / mm is designated as the normal illumination area. 2 The corresponding areas are designated as normal illumination areas. Areas with skin temperature < 37℃ are designated as normal illumination areas, and areas with H < 1mm and D < 3mm are designated as normal illumination areas.

[0044] In one possible implementation, such as Figure 2 As shown, the steps for dividing the first safety buffer zone include: Obtain the boundary contours of the normal irradiation zone and the prohibited irradiation zone on the three-dimensional model of the body surface grading; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with the preset curvature threshold range, and the feature recognition confidence is matched and compared with the preset feature recognition threshold range. Adjust the width of the boundary contour based on the matching comparison results of confidence scores for surface curvature and / or feature identification; The adjusted boundary contour is used as the first safety buffer zone and mapped onto the corresponding area of ​​the body surface graded 3D model.

[0045] In this embodiment, the formula for calculating the curvature of the body surface is as follows: ,in, For the depth gradient, the feature recognition confidence score (γ) is the confidence level output by the feature recognition model that the region belongs to the current security level, and its value ranges from 0 to 1. When K < 0.5 and γ (feature recognition confidence score) ≥ 0.8, the width of the boundary contour is increased by 2mm; when 0.5 ≤ K < 1.2 and γ ≥ 0.8, the width of the boundary contour is increased by 3mm; when K ≥ 1.2 and γ ≥ 0.8, the width of the boundary contour is increased by 4mm; when K < 0.5 and 0.6 ≤ γ < 0.8, the width of the boundary contour is increased by 3mm; when 0.5 ≤ K < 1.2 When 0.6 ≤ γ < 0.8, the width of the boundary profile is increased by 4 mm; when K ≥ 1.2 and 0.6 ≤ γ < 0.8, the width of the boundary profile is increased by 5 mm; when K < 0.5 and γ < 0.6, the width of the boundary profile is increased by 4 mm; when 0.5 ≤ K < 1.2 and γ < 0.6, the width of the boundary profile is increased by 5 mm; when K ≥ 1.2 and γ < 0.6, the width of the boundary profile is increased by 6 mm. Then, the boundary profile with the adjusted width is mapped as the first safety buffer zone to the boundary grid element of the normally lit area and the prohibited area to form a transition area.

[0046] In one possible implementation, the steps for defining the second safety buffer zone include: Obtain the boundary contours of the restricted and prohibited light zones on the three-dimensional model of the body surface classification; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with the preset curvature threshold range, and the feature recognition confidence is matched and compared with the preset feature recognition threshold range. Adjust the width of the boundary contour based on the matching comparison results of confidence scores for surface curvature and / or feature identification; The adjusted boundary contour is used as a second safety buffer zone and mapped onto the corresponding area of ​​the body surface graded 3D model.

[0047] In this embodiment, the adjustment rules for the second safety buffer zone are the same as those for the first safety buffer zone, and will not be repeated here.

[0048] In one possible implementation, the steps of performing the irradiation operation using a fat-dissolving laser also include: The three-dimensional model of body surface grading is divided into several sub-regions; Select non-adjacent sub-regions at preset intervals; Irradiate according to the pattern corresponding to the sub-region; After irradiation, the sub-region is cooled down.

[0049] In this embodiment, sub-regions are divided into 1.8cm × 1.8cm sections. This size avoids excessive local temperature rise caused by an excessively large single irradiation area while ensuring operational efficiency. The preset interval distance is 3mm. Selecting non-adjacent sub-regions reduces heat conduction interference between adjacent areas. Each group selects 6-8 sub-regions for simultaneous irradiation. Cooling is achieved using a contact-type TEC thermoelectric cooler with a transparent quartz cooling window. The cooling power for the normally irradiated sub-region is set to 50-55W, 55-60W for the restricted irradiation area, 60-65W for the first safety buffer zone, and 65-70W for the second safety buffer zone. The target cooling temperature is 37℃, which is monitored in real time by a thermal infrared camera. Once the sub-region temperature drops below 37℃, the next group of sub-regions is irradiated. The entire process is automatically scheduled by a timing control algorithm.

[0050] In one possible implementation, the feature is that during the irradiation operation using the grease-melting laser, the grease-melting safety index data is monitored in real time. If an abnormality is detected in the grease-melting safety index data, the irradiation parameters of the corresponding area are adjusted or the irradiation is immediately paused and a prompt signal is triggered. The safety indicators for fat melting include one or more of the following: skin surface temperature, the fit between the device and the skin surface, feature recognition status, and laser power fluctuation.

[0051] In one possible implementation, real-time monitoring of grease melt safety indicators includes: The skin surface temperature is monitored in real time by a thermal infrared camera. If the skin surface temperature exceeds the preset temperature safety warning value, the irradiation will be stopped immediately and an audio-visual warning will be triggered.

[0052] In this embodiment, when the skin surface temperature exceeds 40°C, the system will pause irradiation of that area within 100ms to prevent skin burns from high temperatures. Simultaneously, an intermittent beeping sound will be emitted via a buzzer. The system will restart once the skin surface temperature drops below 37°C. It should be noted that the preset values ​​described above are typical configurations for this embodiment and can be adjusted within a safe range in practical applications. This does not limit the scope of the invention.

[0053] In one possible implementation, real-time monitoring of grease-melting safety indicators also includes: The device monitors the fit between the device and the skin surface in real time using a pressure sensor. If the fit is lower than the preset fit threshold, the irradiation will be paused immediately and an audio-visual alert will be triggered.

[0054] In this embodiment, the preset pressure threshold is 5N. When the pressure sensor detects a pressure below 5N, it indicates that the device has shifted or is not fully in contact with the target area. In this case, the laser may deviate from its irradiation range due to optical path misalignment, or even mistakenly irradiate non-melting areas. The system will immediately pause irradiation and emit an intermittent beeping sound. The operation will restart once the pressure sensor detects a pressure value above 5N. It should be noted that the above-mentioned preset values ​​are typical configurations for this embodiment and can be adjusted within a safe range in actual applications. This does not limit the invention.

[0055] In one possible implementation, real-time monitoring of grease-melting safety indicators also includes: The safety level identification status of each area is verified by a feature recognition model. If an abnormality in the safety level identification occurs, the irradiation parameters of the corresponding area are reduced, or irradiation is immediately suspended and a warning signal is triggered.

[0056] In this embodiment, the feature recognition model performs verification twice per second, comparing the current recognition result with the initial recognition result. If a safety level change occurs, such as a normal illumination area becoming a restricted illumination area, then irradiation is carried out according to the irradiation parameters corresponding to the restricted illumination area. If a normal illumination area becomes a no-illumination area or a restricted illumination area becomes a no-illumination area, irradiation of the corresponding area is immediately suspended, and an intermittent beeping is emitted through a buzzer.

[0057] In one possible implementation, real-time monitoring of grease-melting safety indicators also includes: The power fluctuation of the laser is monitored in real time by a power sensor. If the power fluctuation exceeds the preset power fluctuation range, the laser output will be stopped immediately and a prompt signal will be triggered.

[0058] In this embodiment, if the laser power fluctuation is greater than ±8%, the system will immediately pause irradiation and trigger an alarm. It should be noted that the above-mentioned preset values ​​are typical configurations for this embodiment; in actual applications, they can be adjusted within a safe range and are not intended to limit the invention.

[0059] In one possible implementation, the degreasing method further includes storing the irradiation parameters, degreasing safety index data, and abnormal handling information for each area in a log table after the irradiation operation is completed.

[0060] In this embodiment, the log table is stored in CSV format, with fields including the ID, security level, mask transmittance, laser output power, laser duty cycle, irradiation duration, skin surface temperature, cooling duration, anomaly type, and anomaly handling measures for each region. The purpose is to leave audit and compliance records to facilitate subsequent traceability and summarization.

[0061] In one possible implementation, the degreasing method also includes: Extract the regions corresponding to security level anomalies; The regions are manually labeled with security levels and then added to the training set of the feature recognition model.

[0062] In this embodiment, the manual annotation process is completed by professionals. After annotation, the data in that area is anonymized to prevent the leakage of user privacy. When added to the training set, it is mixed with the original samples. The model parameters are optimized through incremental training. Typically, the model is updated every 100 anomaly annotation cases to continuously improve the feature recognition model's ability to identify the security level of complex skin, thereby continuously improving the accuracy of subsequent security classification.

[0063] Secondly, such as Figure 3 As shown, this application also provides an automatic skin feature recognition fat-dissolving system for implementing the above-mentioned fat-dissolving method, comprising: The data acquisition unit is configured to collect skin feature data of the area to be melted. The model building unit is configured to build a three-dimensional surface model of the body surface region; The safety level identification unit is configured to obtain a three-dimensional model of the body surface with three safety levels, including normal irradiation zone, limited irradiation zone and prohibited irradiation zone, based on skin feature data and three-dimensional surface model, and to divide a first safety buffer zone between the normal irradiation zone and the prohibited irradiation zone, and a second safety buffer zone between the limited irradiation zone and the prohibited irradiation zone. The laser control unit is configured to irradiate the normal illumination area, the limited illumination area, the first safety buffer zone, and the second safety buffer zone using the first to fourth modes respectively, and to shield the forbidden illumination area from irradiation. The energy output of the first to fourth modes decreases step by step within a unit time. The laser emitting unit is configured to irradiate the melting area according to the laser irradiation parameters; The cooling unit is configured to cool the melting area during the melting process.

[0064] In one possible implementation, the grease-dissolving system also includes: The safety indicator monitoring unit is configured to monitor the grease melting safety indicator data in real time during the irradiation operation. If abnormalities are detected in the grease melting safety indicator data, the laser irradiation parameters of the corresponding area will be adjusted or the irradiation will be immediately suspended and a warning signal will be triggered.

[0065] Thirdly, this application also provides a storage medium for automatically identifying skin features, wherein the storage medium stores instructions that, when invoked by a processor, are used to execute any of the above-mentioned fat-dissolving methods.

[0066] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatically identifying skin features in fat reduction, characterized in that, Includes the following steps: Collect skin feature data and construct a three-dimensional surface model of the body surface area; The skin feature data and three-dimensional surface model are input into the pre-trained feature recognition model. The feature recognition model classifies the skin region according to the safety level, and obtains a three-dimensional model of the body surface graded into normal irradiation area, limited irradiation area and prohibited irradiation area. Based on the depth gradient of the three-dimensional model of the body surface grading and the feature recognition confidence of the feature recognition model, a first safety buffer zone is divided into the normal irradiation zone and the no-irradiation zone, and a second safety buffer zone is divided into the limited irradiation zone and the no-irradiation zone. Based on the aforementioned three-dimensional model of body surface grading, an irradiation plan is generated; According to the irradiation plan, irradiation operations are performed, wherein the irradiation plan includes: irradiating the normally lit area according to a first mode, irradiating the restricted area according to a second mode; irradiating the first safety buffer zone according to a third mode, irradiating the second safety buffer zone according to a fourth mode, and shielding the prohibited area with irradiation. The energy output per unit time decreases progressively in the first, second, third, and fourth modes.

2. The fat-dissolving method for automatically identifying skin features according to claim 1, characterized in that, The skin feature data includes one or more of the following: skin appearance data, deep skin structure data, skin temperature data, and three-dimensional morphology data of the skin surface.

3. The fat-dissolving method for automatically identifying skin features according to claim 2, characterized in that, The steps to obtain the three-dimensional model of the body surface grading include: The skin feature data and three-dimensional surface model are input into the feature recognition model; The feature recognition model marks the areas corresponding to skin appearance data that meet the preset high-risk standards as no-photography zones. The feature recognition model marks the areas corresponding to the deep skin structure data that meet the preset high-risk standards as forbidden areas for exposure. The feature recognition model identifies areas corresponding to skin temperature data that meet preset high-risk standards as no-sunlight zones. The feature recognition model identifies areas corresponding to the three-dimensional morphological data of the skin surface that meet the preset high-risk standards as no-photography zones.

4. The fat-dissolving method for automatically identifying skin features according to claim 3, characterized in that, The steps for obtaining the three-dimensional model of the body surface grading also include: The feature recognition model identifies the areas corresponding to skin appearance data that meet the preset medium-risk standards as restricted exposure areas. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset medium-risk standards as restricted exposure areas. The feature recognition model identifies the areas corresponding to skin temperature data that meet the preset medium-risk standards as restricted-sunlight zones. The feature recognition model identifies the areas corresponding to the three-dimensional morphological data of the skin surface that meet the preset medium-risk standards as restricted exposure areas.

5. The fat-dissolving method for automatically identifying skin features according to claim 3, characterized in that, The steps for obtaining the three-dimensional model of the body surface grading also include: The feature recognition model identifies the areas corresponding to skin appearance data that meet the preset low-risk standard as normal exposure areas. The feature recognition model identifies the areas corresponding to the deep skin structure data that meet the preset low-risk criteria as normal exposure areas. The feature recognition model identifies the area corresponding to skin temperature data that meets the preset low-risk standard as the normal irradiation area. The feature recognition model identifies the area corresponding to the three-dimensional morphological data of the skin surface that meets the preset low-risk standard as the normal irradiation area. The feature recognition model maps the normal illumination area, limited illumination area, and no-illumination area to the corresponding regions of the three-dimensional surface model, respectively, to obtain a graded three-dimensional model of the body surface.

6. The fat-dissolving method for automatically identifying skin features according to claim 5, characterized in that, The steps for defining the first safety buffer zone include: Obtain the boundary contours of the normally illuminated area and the shaded area on the three-dimensional model of the body surface classification; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the feature recognition confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with a preset curvature threshold range, and the feature recognition confidence is matched and compared with a preset feature recognition threshold range. The width of the boundary contour is adjusted based on the matching comparison results of the confidence scores of the surface curvature and / or feature identification. The adjusted boundary contour is used as the first safety buffer zone and mapped onto the corresponding area of ​​the body surface graded three-dimensional model.

7. The fat-dissolving method for automatically identifying skin features according to claim 5, characterized in that, The steps for defining the second safety buffer zone include: Obtain the boundary contours of the restricted and prohibited light zones on the three-dimensional model of the body surface grading; Extract the depth gradient of the boundary contour, and calculate the surface curvature based on the depth gradient; Obtain the feature recognition confidence level of the feature recognition model for the boundary contour; The surface curvature is matched and compared with a preset curvature threshold range, and the feature recognition confidence is matched and compared with a preset feature recognition threshold range. The width of the boundary contour is adjusted based on the matching comparison results of the confidence scores of the surface curvature and / or feature identification. The adjusted boundary contour is used as a second safety buffer zone and mapped onto the corresponding area of ​​the body surface graded three-dimensional model.

8. The fat-dissolving method for automatically identifying skin features according to claim 1, characterized in that, The steps involved in performing irradiation using a fat-dissolving laser also include: The three-dimensional model of body surface grading is divided into several sub-regions; Select non-adjacent sub-regions at preset intervals; Irradiate according to the pattern corresponding to the sub-region; After irradiation, the sub-region is cooled down.

9. A method for automatically identifying skin features for fat reduction according to any one of claims 1 to 8, characterized in that, During the irradiation operation using the degreasing laser, the degreasing safety index data is monitored in real time. If abnormalities are detected in the degreasing safety index data, the irradiation parameters of the corresponding area are adjusted or the irradiation is immediately paused and a warning signal is triggered. The fat melting safety index data includes one or more of the following: skin surface temperature, the fit between the device and the skin surface, feature recognition status, and laser power fluctuation.

10. The fat-dissolving method for automatically identifying skin features according to claim 9, characterized in that, Real-time monitoring data on grease melting safety indicators includes: The skin surface temperature is monitored in real time by a thermal infrared camera. If the skin surface temperature exceeds the preset temperature safety warning value, the irradiation will be stopped immediately and an audio-visual warning will be triggered.

11. The fat-dissolving method for automatically identifying skin features according to claim 9, characterized in that, Real-time monitoring of fat melting safety indicators also includes: The device monitors the fit between the device and the skin surface in real time using a pressure sensor. If the fit is lower than the preset fit threshold, the irradiation will be paused immediately and an audio-visual alert will be triggered.

12. The fat-dissolving method for automatically identifying skin features according to claim 9, characterized in that, Real-time monitoring of fat melting safety indicators also includes: The safety level identification status of each area is verified by a feature recognition model. If an abnormality in the safety level identification occurs, the irradiation parameters of the corresponding area are reduced, or irradiation is immediately suspended and a warning signal is triggered.

13. The fat-dissolving method for automatically identifying skin features according to claim 9, characterized in that, Real-time monitoring of fat melting safety indicators also includes: The power fluctuation of the laser is monitored in real time by a power sensor. If the power fluctuation exceeds the preset power fluctuation range, the laser output will be stopped immediately and a prompt signal will be triggered.

14. The fat-dissolving method for automatically identifying skin features according to claim 9, characterized in that, The degreasing method also includes: after the irradiation operation is completed, storing the irradiation parameters, degreasing safety index data and abnormal handling information of each area in a log table.

15. The fat-dissolving method for automatically identifying skin features according to claim 12, characterized in that, The fat melting method also includes: Extract the regions corresponding to security level anomalies; The region is manually labeled with a security level, and then added to the training set of the feature recognition model.

16. A fat-dissolving system that automatically identifies skin features, characterized in that, include: The data acquisition unit is configured to collect skin feature data of the area to be melted. The model building unit is configured to build a three-dimensional surface model of the body surface region; The safety level identification unit is configured to obtain a three-dimensional model of the body surface with three safety levels, including normal irradiation zone, limited irradiation zone and prohibited irradiation zone, based on skin feature data and three-dimensional surface model, and to divide a first safety buffer zone between the normal irradiation zone and the prohibited irradiation zone, and a second safety buffer zone between the limited irradiation zone and the prohibited irradiation zone. The laser control unit is configured to irradiate the normal illumination area, the limited illumination area, the first safety buffer zone, and the second safety buffer zone using the first to fourth modes respectively, and to shield the forbidden illumination area from irradiation. The energy output of the first to fourth modes decreases step by step within a unit time. The laser emitting unit is configured to irradiate the melting area according to the laser irradiation parameters; The cooling unit is configured to cool the melting area during the melting process.

17. The fat-dissolving system for automatically recognizing skin features according to claim 16, characterized in that, The fat melting system also includes: The safety indicator monitoring unit is configured to monitor the grease melting safety indicator data in real time during the irradiation operation. If abnormalities are detected in the grease melting safety indicator data, the laser irradiation parameters of the corresponding area will be adjusted or the irradiation will be immediately suspended and a warning signal will be triggered.

18. A storage medium for automatically identifying skin features, characterized in that, The storage medium stores instructions that, when invoked by a processor, are used to execute the grease-melting method according to any one of claims 1 to 16.

Images