CN-122005073-A - Fat melting method, system and storage medium for automatically identifying skin characteristics

Abstract

The invention discloses a fat melting method, a fat melting system and a storage medium for automatically identifying skin features, wherein the fat melting method comprises the steps of collecting skin feature data, constructing a three-dimensional surface model of a body surface area, inputting the skin feature data and the three-dimensional surface model into a feature identification model which is pre-trained, classifying the skin area by the feature identification model to obtain a body surface hierarchical three-dimensional model comprising a normal illumination area, a limited illumination area and a forbidden illumination area, dividing the normal illumination area and the forbidden illumination area into a first safety buffer zone and dividing the limited illumination area and the forbidden illumination area into a second safety buffer zone based on the depth gradient of the body surface hierarchical three-dimensional model and the feature identification confidence of the feature identification model, generating an illumination scheme according to the body surface hierarchical three-dimensional model, and executing illumination operation according to the illumination scheme.

Inventors

• TANG CHAOHUI

Assignees

• 深圳市桥福智能设备有限公司

Dates

Publication Date

20260512

Application Date

20251231

Claims (18)

1. 1. A method for automatically identifying skin characteristics, comprising the steps of: collecting skin characteristic data and constructing a three-dimensional surface model of a body surface area; inputting the skin characteristic data and the three-dimensional surface model into a pre-trained characteristic recognition model, and classifying the skin area by the characteristic recognition model in a safety level manner to obtain a body surface hierarchical three-dimensional model comprising a normal illumination area, a limited illumination area and a forbidden illumination area; Dividing a first safety buffer zone for the normal illumination zone and the forbidden illumination zone and dividing a second safety buffer zone for the limited illumination zone and the forbidden illumination zone based on the depth gradient of the body surface grading three-dimensional model and the feature recognition confidence of the feature recognition model; Generating an irradiation scheme according to the body surface grading three-dimensional model; According to the irradiation scheme, the irradiation operation is executed, wherein the irradiation scheme comprises the steps of irradiating a normal irradiation area according to a first mode, irradiating a limited irradiation area according to a second mode, irradiating a first safety buffer belt according to a third mode, irradiating a second safety buffer belt according to a fourth mode, and shielding irradiation of a forbidden irradiation area; the energy output in the first, second, third and fourth modes is gradually reduced in a unit time.
2. 2. The method of claim 1, wherein the skin characteristic data comprises one or more of skin appearance data, skin deep structure data, skin temperature data, and skin surface three-dimensional topography data.
3. 3. The method for automatically identifying skin features of claim 2, wherein the step of obtaining the body surface hierarchical three-dimensional model comprises: inputting the skin feature data and the three-dimensional surface model into a feature recognition model; The feature recognition model marks a region corresponding to the skin appearance data meeting a preset high risk standard as a forbidden region; The feature recognition model marks a region corresponding to the skin deep structure data meeting a preset high risk standard as a forbidden region; the feature recognition model marks a region corresponding to skin temperature data meeting a preset high risk standard as a forbidden region; And the feature recognition model marks the area corresponding to the three-dimensional shape data of the skin surface, which accords with a preset high risk standard, as a forbidden area.
4. 4. A method of automatically identifying skin features as in claim 3 wherein the step of obtaining the body surface hierarchical three-dimensional model further comprises: the feature recognition model marks the area corresponding to the skin appearance data which accords with the preset risk standard as a limiting area; The feature recognition model marks a region corresponding to the skin deep structure data which accords with a preset risk standard to be a limiting region; the feature recognition model marks a region corresponding to skin temperature data meeting a preset risk criterion to be a limiting region; the feature recognition model marks the area corresponding to the three-dimensional shape data of the skin surface, which accords with the preset risk standard, as a limiting area.
5. 5. A method of automatically identifying skin features as in claim 3 wherein the step of obtaining the body surface hierarchical three-dimensional model further comprises: the feature recognition model marks an area corresponding to the skin appearance data which accords with a preset low risk standard as a normal shot area; The feature recognition model marks an area corresponding to the skin deep structure data meeting a preset low risk standard as a normal area; the feature recognition model marks an area corresponding to skin temperature data meeting a preset low risk standard as a normal shot area; the feature recognition model marks an area corresponding to the three-dimensional shape data of the skin surface, which accords with a preset low risk standard, as a normal area; And mapping the normal shot region, the limited shot region and the forbidden shot region to corresponding regions of the three-dimensional surface model by the feature recognition model respectively to obtain a body surface grading three-dimensional model.
6. 6. The method of claim 5, wherein the step of dividing the first safety buffer band comprises: obtaining boundary contours of the normal illumination area and the forbidden illumination area on the body surface hierarchical three-dimensional model; extracting depth gradient of the boundary contour, and calculating to obtain body surface curvature based on the depth gradient; acquiring the feature recognition confidence coefficient of the feature recognition model on the boundary contour; Matching and matching the body surface curvature with a preset curvature threshold range, and matching the feature recognition confidence coefficient with a preset feature recognition threshold range; According to the matching comparison result of the body surface curvature and/or the feature recognition confidence coefficient, the width of the boundary contour is adjusted; And mapping the adjusted boundary profile serving as a first safety buffer zone to a corresponding region of the body surface grading three-dimensional model.
7. 7. The method of claim 5, wherein the step of dividing the second safety buffer band comprises: obtaining boundary contours of the limiting area and the forbidden area on the body surface hierarchical three-dimensional model; extracting depth gradient of the boundary contour, and calculating to obtain body surface curvature based on the depth gradient; acquiring the feature recognition confidence coefficient of the feature recognition model on the boundary contour; Matching and matching the body surface curvature with a preset curvature threshold range, and matching the feature recognition confidence coefficient with a preset feature recognition threshold range; According to the matching comparison result of the body surface curvature and/or the feature recognition confidence coefficient, the width of the boundary contour is adjusted; and mapping the adjusted boundary profile serving as a second safety buffer zone to a corresponding region of the body surface grading three-dimensional model.
8. 8. The method for automatically identifying skin features according to claim 1, wherein the step of performing the irradiation operation using the fat-melting laser further comprises: Dividing the body surface grading three-dimensional model into a plurality of subareas; selecting non-adjacent subareas according to a preset interval distance; Irradiating according to the mode corresponding to the subarea; And after the irradiation is finished, cooling the sub-area.
9. 9. The method for automatically identifying skin features according to any one of claims 1 to 8, wherein in the process of performing irradiation operation using a fat-melting laser, the fat-melting safety index data is monitored in real time, and if abnormality is detected in the fat-melting safety index data, irradiation parameters of the corresponding region are adjusted or irradiation is immediately suspended and a prompt signal is triggered; The fat melting safety index data comprise one or more of skin surface temperature, fitting degree of equipment and skin surface, characteristic identification state and laser power fluctuation.
10. 10. The method for automatically identifying skin features according to claim 9, wherein monitoring the safety index data of the molten fat in real time comprises: The skin temperature is monitored in real time through the thermal infrared camera, and if the skin temperature exceeds a preset temperature safety early warning value, the irradiation is immediately paused and an audible and visual prompt is touched.
11. 11. The method for automatically identifying skin features according to claim 9, wherein monitoring the safety index data of the molten fat in real time further comprises: The bonding degree of the equipment and the skin surface is monitored in real time through the pressure sensor, and if the bonding degree is lower than a preset bonding threshold value, the irradiation is immediately suspended and the audible and visual prompt is touched.
12. 12. The method for automatically identifying skin features according to claim 9, wherein monitoring the safety index data of the molten fat in real time further comprises: And verifying the safety level identification state of each region through the characteristic identification model, and if the safety level identification abnormality occurs, reducing the irradiation parameters of the corresponding region, or immediately suspending irradiation and triggering a prompt signal.
13. 13. The method for automatically identifying skin features according to claim 9, wherein monitoring the safety index data of the molten fat in real time further comprises: and monitoring the power fluctuation of the laser in real time through a power sensor, and immediately stopping the laser output and triggering a prompt signal if the power fluctuation exceeds a preset power fluctuation range.
14. 14. The method for automatically identifying skin features according to claim 9, wherein the method further comprises storing the irradiation parameters, the safety index data of the molten fat and the abnormal conditions of each region in a log table after the irradiation operation is completed.
15. 15. The method for automatically identifying skin features according to claim 12, wherein said method for thawing further comprises: extracting a region corresponding to the security level identification abnormality; and marking the safety level of the region by manpower, and adding the region into a training set of the feature recognition model.
16. 16. A fat melting system for automatically identifying skin features, comprising: a data acquisition unit configured to acquire skin characteristic data of a region to be fat-melted; A model building unit configured to build a three-dimensional surface model of the body surface region; The safety level identification unit is configured to obtain a body surface grading three-dimensional model of three safety levels comprising a normal shot region, a limited shot region and a forbidden shot region according to the skin characteristic data and the three-dimensional surface model, a first safety buffer zone is divided between the normal shot region and the forbidden shot region, and a second safety buffer zone is divided between the limited shot region and the forbidden shot region; the laser regulation and control unit is configured to irradiate the normal irradiation area, the limited irradiation area, the first safety buffer belt and the second safety buffer belt in a first mode to a fourth mode respectively, shield the irradiation area from irradiation, and gradually reduce the energy output by the first mode to the fourth mode in unit time; a laser emission unit configured to irradiate the fat melting region according to laser irradiation parameters; And the cooling unit is configured to cool down the fat melting area in the fat melting operation process.
17. 17. The fat melting system for automatically identifying skin features according to claim 16, further comprising: The safety index monitoring unit is configured to monitor the molten fat safety index data in real time in the irradiation process, and if abnormal molten fat safety index data is monitored, the laser irradiation parameters of the corresponding area are adjusted or irradiation is immediately suspended and a prompt signal is triggered.
18. 18. A storage medium having stored therein instructions for executing the fat melting method according to any one of claims 1 to 16 when called by a processor.

Description

Fat melting method, system and storage medium for automatically identifying skin characteristics Technical Field The invention relates to the technical field of non-invasive human tissue heating/fat reduction, belongs to the optical and control technology of medical/household beauty equipment, and particularly relates to a fat melting method, system and storage medium capable of automatically identifying skin characteristics. Background With the increasing use of non-invasive lipid melting technology in the cosmetic and health fields, the demands of users for their automation and therapeutic safety continue to increase. However, the existing fat melting scheme is mostly dependent on experience of operators to judge and remove skin sensitive areas, so that subjectivity is high, identification repeatability among different operators is poor, normal areas and sensitive areas cannot be accurately distinguished, misirradiation of laser on the sensitive areas is easily caused, and further local overheating damage of the sensitive areas is caused, meanwhile, although the existing fat melting technology is used for carrying out regional fat melting treatment on the normal areas and the sensitive areas of a human body, safety transition belts are not arranged between the normal areas and the sensitive areas, so that energy is not smoothly transited when the laser passes through the sensitive areas from the normal areas, and further edge parts of the sensitive areas are scalded, or fat melting effect of edges of the normal areas is not good. Therefore, designing a method, a system and a storage medium for automatically identifying skin characteristics, accurately dividing safety level areas and connecting a regulation and control irradiation scheme becomes a problem to be solved in the field. Disclosure of Invention The invention is realized mainly by the following technical scheme: the invention provides a fat melting method capable of automatically identifying skin characteristics, which comprises the following steps: collecting skin characteristic data and constructing a three-dimensional surface model of a body surface area; inputting the skin characteristic data and the three-dimensional surface model into a pre-trained characteristic recognition model, and classifying the safety level of the skin area by the characteristic recognition model to obtain a body surface hierarchical three-dimensional model comprising a normal illumination area, a limited illumination area and a forbidden illumination area; based on the depth gradient of the body surface grading three-dimensional model and the feature recognition confidence of the feature recognition model, dividing a first safety buffer zone into a normal illumination zone and a forbidden illumination zone, and dividing a second safety buffer zone into a limited illumination zone and a forbidden illumination zone; Generating an irradiation scheme according to the body surface grading three-dimensional model; according to an irradiation scheme, performing irradiation operation, wherein the irradiation scheme comprises the steps of irradiating a normal irradiation region according to a first mode, irradiating a limited irradiation region according to a second mode, irradiating a first safety buffer belt according to a third mode, irradiating a second safety buffer belt according to a fourth mode, and shielding irradiation of the forbidden irradiation region; The energy output in the first, second, third and fourth modes is stepped down in a unit time. Further, the skin characteristic data includes one or more of skin appearance data, skin deep structure data, skin temperature data, and skin surface three-dimensional topography data. Further, the step of obtaining the body surface grading three-dimensional model comprises the following steps: inputting the skin characteristic data and the three-dimensional surface model into a characteristic recognition model; the feature recognition model marks the area corresponding to the skin appearance data meeting the preset high risk standard as a forbidden area; The feature recognition model marks a region corresponding to the skin deep structure data meeting a preset high risk standard as a forbidden region; The feature recognition model marks the area corresponding to the skin temperature data meeting the preset high risk standard as a forbidden area; the feature recognition model marks the area corresponding to the three-dimensional shape data of the skin surface, which accords with a preset high risk standard, as a forbidden area. Further, the step of obtaining the body surface grading three-dimensional model further comprises the steps of: The feature recognition model marks the area corresponding to the skin appearance data which accords with the preset risk standard as a limiting area; the feature recognition model marks the region corresponding to the skin deep structure data which accords with the preset risk standard