Facial acupuncture curative effect comprehensive evaluation method based on multi-angle optical imaging

Abstract

The invention discloses a facial acupuncture curative effect comprehensive evaluation method based on multi-angle optical imaging, and the method comprises the steps: extracting a blood flow velocity dynamic signal of a subcutaneous capillary, extracting a human face qi and blood ruddiness index and a body fluid glossiness index, and obtaining a wrinkle center line and three-dimensional wrinkle form parameters. Establishing a correlation coefficient between the exercise intensity and the acupoint distance; and constructing an acupuncture curative effect evaluation data set, matching the acupuncture curative effect evaluation data set with a traditional Chinese medicine syndrome type set threshold, and fusing multi-modal data by adopting a deep learning model to generate a quantitative evaluation result. Synchronous acquisition of multi-dimensional information such as complexion, qi and blood, skin structure, muscle dynamics and the like is realized through an integrated system, complex operation and high cost caused by multi-sensor fusion are avoided, the defect of insufficient data dimension of single equipment is made up, and the system is adaptive to practical clinical application scenes of traditional Chinese medicine.

Description

Technical Field

[0001] This invention relates to the fields of medical testing and traditional Chinese medicine engineering technology, and in particular to a comprehensive evaluation method for the efficacy of facial acupuncture based on multi-angle optical imaging. Background Technology

[0002] The evaluation of the efficacy of facial acupuncture therapy needs to take into account facial complexion, muscle function, and skin morphology. Existing technologies are either too inaccurate, like ordinary cameras, or too complex and expensive, like hyperspectral and 3D motion capture systems. The market lacks a dedicated device that can provide reliable quantitative data while being easy to operate. Systems based on a single color camera are low in cost, but have limited evaluation dimensions and accuracy.

[0003] Existing facial assessment technologies have significant limitations: one type is conventional imaging equipment, which, although easy to operate and low in cost, can only acquire two-dimensional surface information and cannot distinguish between skin surface reflection and subcutaneous physiological signals, making it difficult to accurately reflect the state of Qi and blood and changes in deep tissues; the other type is high-end multi-sensor systems, which can achieve multi-dimensional data acquisition, but the equipment structure is complex, the operation is cumbersome and the cost is high, and the requirements for the use environment and the professional skills of the operators are strict, making it difficult to adapt to the actual application scenarios of TCM clinical practice.

[0004] Meanwhile, the evaluation of the efficacy of acupuncture in traditional Chinese medicine needs to take into account the coordinated changes of qi and blood, meridians, and morphology. Existing technologies often lack targeted adaptation to traditional Chinese medicine theories, either focusing only on a single physiological indicator or deviating from the core of meridian differentiation, resulting in a weak correlation between the collected data and the clinical efficacy judgment, and failing to provide effective support for the diagnosis and treatment of traditional Chinese medicine. Summary of the Invention

[0005] The purpose of this invention is to address the technical deficiencies in the existing technology by providing a comprehensive evaluation method for the therapeutic effect of facial acupuncture based on multi-angle optical imaging.

[0006] The technical solution adopted to achieve the purpose of this invention is: A comprehensive evaluation method for the therapeutic effect of facial acupuncture based on multi-angle optical imaging includes the following steps: Step 1: Scan the human face and, in conjunction with the facial meridian map, designate the acupoints, meridians, and the area within 1.5cm around the acupuncture points on the face as the evaluation area. ; Step 2: Continuously capture images of the human face, select clear images, and calibrate the ambient light intensity in real time. Then, extract the dynamic signal of blood flow velocity in subcutaneous capillaries using a polarization extinction model and spectral separation algorithm. ; Step 3: Extract the facial rosy complexion index based on the CIELab color space. And the gloss index of body fluids ; Step 4: Acquire striped projection images, perform 3D reconstruction, generate a two-layer 3D model, and overlay facial meridian lines. and assessment area Obtain the center line of the wrinkle and three-dimensional wrinkle morphology parameters ; Step 5: Guide the execution of the set facial expression sequence, capture the instantaneous changes in facial muscle movement, complete muscle dynamics acquisition, and obtain the angle between the muscle movement direction of each pixel and the meridian direction. Furthermore, a correlation coefficient between exercise intensity and acupoint distance was established within the assessment area. ; Step 6, the blood flow velocity dynamic signal from Step 2 The blood and complexion index obtained in step 3 Surface gloss index The three-dimensional wrinkle morphology parameters obtained in step 4 Step 5 obtains the angle between the muscle movement direction of the pixel and the meridian direction. The correlation coefficient between muscle movement intensity and acupoint distance within the region was assessed. Construct a dataset for evaluating the efficacy of acupuncture. The acupuncture efficacy evaluation dataset was matched with thresholds set for TCM syndrome types, and a deep learning model was used to fuse multimodal data to generate evaluation results.

[0007] In the above technical solution, continuous image capture is performed using the following process: A feature spectrum with three different polarization angles (0°, 45°, and 90°) is sequentially output through an LED array. This feature spectrum includes four types of light: 450nm blue light, 550nm green light, 650nm red light, and 940nm near-infrared light, generating 12 sets of imaging conditions. Under each imaging condition, the human face is continuously captured by a camera to obtain continuous images. Clear images are selected from the continuous images under each imaging condition based on the gradient value of each pixel on the continuous image plane. The ambient light intensity of the clear images under each imaging condition is recorded, and the ambient light intensity is calibrated in real time using a light sensor. The real-time calibrated clear images are then processed using a polarization extinction model and a spectral separation algorithm to highlight the oil reflection on the facial skin surface and extract the dynamic signal of blood flow velocity in subcutaneous capillaries.

[0008] In the above technical solution, the gradient value selection is achieved through the following formula: ;in, , The Sobel operator is used to compute the images in... Pixelx axis, y Axial gradient value, For overall image sharpness, The sharpness threshold is used; if the formula is satisfied, the image is considered sharp. The real-time calibration is achieved through the following formula: ;in, The brightness values ​​of the original, clear image captured by the camera. For standard lighting intensity, This refers to the calibrated image brightness value.

[0009] In the above technical solution, the polarization extinction model is: ,in, For clear images calibrated in real time, The intensity of mirror-like reflection on the skin surface. For extinction efficiency, To eliminate the effective light intensity in the subcutaneous tissue after reflection; The dynamic signal of blood flow velocity in subcutaneous capillaries was extracted using the following blood flow velocity model. ,in, This represents the dynamic signal of blood flow velocity in subcutaneous capillaries. For calibration coefficients, This represents the change in light intensity under near-infrared conditions. For the time of collection, This refers to the effective subcutaneous light intensity in the near-infrared band.

[0010] In the above technical solution, in step 3, based on the CIELab color space and combined with blood oxygen saturation data in the 940nm near-infrared band under imaging conditions, the traditional red-green axis is corrected as the quantification of the blood rosiness index, and the body fluid luster index is inverted through the 450nm blue light band. The formula for extracting the blood rosiness index is as follows: in, The index represents the degree of rosy complexion and health. These are the weighting coefficients. These are the red and green axis values ​​in the CIELab space. The blood oxygen saturation was obtained under 940nm near-infrared imaging conditions; the formula for extracting the body fluid luster index is: in, The gloss index of body fluids. These are the weighting coefficients. The brightness value for the CIELab space. The skin moisture content was obtained by inversion under imaging conditions of 450nm blue light band.

[0011] In the above technical solution, in step 4, the camera and the grating projector are combined to obtain a striped projection image. The striped projection image includes a frontal image, a left 45° facial image, a right 45° facial image, and a combined image with a custom tilt angle. The combined image with the custom tilt angle includes an image of a dense facial wrinkle area and multiple tilt angles of a facial acupuncture point area. The image of a dense facial wrinkle area is the main direction angle of the dense wrinkle area extracted by an edge detection algorithm.

[0012] In the above technical solution, in step 4, the striped projection image is reconstructed in three dimensions, and the macroscopic facial contour is obtained through structured light. And extracting micro-wrinkles using phase deflection technique Blend to generate a two-layer 3D model ; The structured light acquisition of the macroscopic facial contour is calculated using the following formula: ,in, In the striped projection image Macroscopic contour depth of a pixel. For the fringe amplitude, The phase value of the structured light stripe. The stripe spacing; The phase deflection technique extracts micro-wrinkles using the following formula: ,in, for The depth of micro-wrinkles at the pixel level. These are the phase-to-depth conversion coefficients. This represents the phase deflection.

[0013] In the above technical solution, in step 4, facial meridian lines are superimposed on the double-layered three-dimensional model. Obtain the center line of the wrinkle Identify the center line of wrinkles With meridian lines The intersection points are superimposed on the evaluation areas calibrated in step 1 onto the double-layered 3D model. Identify the depth of skin depressions and protrusions around acupoints, and record three-dimensional wrinkle morphology parameters. ; Identify the intersection of wrinkles and meridians using the following formula: and ,in The wrinkle center line is obtained by superimposing meridian lines on a 3D model. For meridian lines, The average depth within a 5mm radius around the intersection. The concave threshold is used; if the formula is satisfied, it is determined to be a valid intersection point. The following formula is used to identify the skin depressions and protrusions around acupoints and record the three-dimensional wrinkle morphology parameters. : ,in, The depression around the acupoint skin in the i-th evaluation region represents the three-dimensional wrinkle morphology parameter. To assess the average depth of the area, The minimum depth value for the core assessment area. For the assessment area.

[0014] In the above technical solution, in step 5, the face is illuminated by a uniform near-infrared light source to avoid strong light stimulation of facial muscles. Then, a set sequence of facial expressions is guided and executed. The set sequence of facial expressions includes: frowning and raising the forehead to activate the Du meridian and Bladder meridian in the forehead; pursing the lips and puffing out the cheeks to activate the Ren meridian and Stomach meridian around the mouth; lifting the side of the face and lifting the corners of the mouth diagonally upward to activate the Gallbladder meridian and Large Intestine meridian in the cheek; blinking and frowning to simultaneously activate the Liver meridian and Triple Energizer meridian around the eyes. The camera captures instantaneous changes in muscle movement during the execution of a pre-defined sequence of facial expressions, records the motion trajectory of the evaluation area, obtains video frames, analyzes pixel motion between video frames, calculates the strain value and contraction velocity of the muscles in the evaluation area, and maps the strain field to the corresponding meridian lines, as well as the contraction velocity to the corresponding meridian lines. Obtain the angle between the muscle movement direction of a pixel and the meridian direction. Furthermore, the correlation between exercise intensity and acupoint distance was established within the assessment area to complete muscle dynamics data acquisition. The analysis of pixel motion between video frames specifically involves calculating the strain values ​​of the muscle in the evaluation region of the pixel at any given time in the x and y axes based on the displacement of the pixel motion between video frames, and then calculating the strain field using the dense optical flow method. ; in , Let be the strain values ​​of the pixel at time t in the x-axis and y-axis directions. , Let x and y be the displacements of the pixel at time t along the x and y axes. , The pixel interval is [number].

[0015] In the above technical solution, the angle between the muscle movement direction of the pixel and the meridian direction, the strain value, and the contraction speed are synchronously mapped to the meridian line. ,include: The angle between the direction of muscle movement at a pixel and the direction of the meridian is calculated using the following formula: ,in For pixels The angle between the direction of muscle movement and the direction of the meridian. The direction vector of muscle movement is mapped from the strain field to the corresponding meridian lines. get, The contraction speed is obtained by mapping the direction vector of the meridian to the corresponding meridian line; Then establish the correlation between exercise intensity and acupoint distance: , Let be the correlation coefficient between muscle movement intensity and acupoint distance in the i-th assessment region. For covariance, For variance, Let (x, y) be the distance from the (x, y) pixel to the center of the i-th acupoint. The intensity of muscle movement at that pixel, i.e. The amplitude.

[0016] Compared with the prior art, the beneficial effects of the present invention are: By deeply integrating the TCM meridian differentiation theory with optical imaging technology, the system can accurately locate acupoints and meridian-related areas, capture the synergistic changes in the flow of Qi and blood, muscle movement and meridian response, and transform the traditional subjective judgment of efficacy into an objective and quantifiable indicator system, which significantly improves the scientificity and credibility of the evaluation results. The integrated system enables the simultaneous collection of multi-dimensional information such as complexion, blood circulation, skin structure, and muscle dynamics. This avoids the complex operation and high cost of multi-sensor fusion, and makes up for the lack of data dimensions of a single device, making it suitable for the actual application scenarios of TCM clinical practice. By utilizing technologies such as dynamic polarization spectroscopy and multi-angle three-dimensional reconstruction, interference signals on the skin surface can be effectively eliminated, and subtle changes in subcutaneous blood and qi dynamics, skin microstructure, and muscle movement can be deeply captured. This allows for the discovery of efficacy differences that are difficult to detect using traditional assessment methods, providing a precise basis for real-time adjustments to acupuncture plans. In line with the dialectical logic of traditional Chinese medicine regarding qi, blood, meridians, and morphology, optical imaging data is combined with traditional Chinese medicine syndrome classification and meridian response patterns. The resulting assessment report contains both objective quantitative data and aligns with the clinical diagnostic thinking of traditional Chinese medicine, providing a feasible path for the assessment of integrated traditional Chinese and Western medicine. Attached Figure Description

[0017] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] like Figure 1 As shown, a comprehensive evaluation method for facial acupuncture efficacy based on multi-angle optical imaging includes the following steps: Step 1: Scan the human face and, in conjunction with the facial meridian map, designate the acupoints, meridians, and the area within 1.5cm around the acupuncture points on the face as the evaluation area. .

[0020] Step 2: Continuously capture images of the human face, select clear images, and calibrate the ambient light intensity in real time. Then, extract the dynamic signal of blood flow velocity in subcutaneous capillaries using a polarization extinction model and spectral separation algorithm. Specifically, it includes the following steps.

[0021] Step 2.1: The LED array sequentially outputs characteristic spectra with three different polarization angles: 0°, 45°, and 90°. The characteristic spectra include four types of light: 450nm blue light, 550nm green light, 650nm red light, and 940nm near-infrared light, generating 12 sets of imaging conditions. Under each set of imaging conditions, the camera continuously captures images of the human face to obtain continuous images.

[0022] Step 2.2: Filter the clear images in each batch of images under each imaging condition by using the gradient value of each pixel on the image plane in Step 2.1, record the ambient light intensity of the clear images under each imaging condition, and calibrate the ambient light intensity in real time using a light sensor. The gradient value filtering is achieved through the following formula: ;in, , The Sobel operator is used to compute the images in... Pixel x axis, y Axial gradient value, For overall image sharpness, The sharpness threshold is used; if the formula is satisfied, the image is considered sharp. The real-time calibration is achieved through the following formula: ;in, The brightness values ​​of the original, clear image captured by the camera. For standard lighting intensity, This refers to the calibrated image brightness value.

[0023] Step 2.3: The clear image obtained in step 2.2 after real-time calibration is used to highlight the oil reflection on the surface of the facial skin and extract the dynamic signal of blood flow velocity of subcutaneous capillaries through the polarization extinction model and spectral separation algorithm. The polarization extinction model: ,in, For clear images calibrated in real time, The intensity of mirror-like reflection on the skin surface. For extinction efficiency, To eliminate the effective light intensity in the subcutaneous tissue after reflection; The dynamic signal of blood flow velocity in subcutaneous capillaries was extracted using the following blood flow velocity model. ,in, This represents the dynamic signal of blood flow velocity in subcutaneous capillaries. For calibration coefficients, This represents the change in light intensity under near-infrared conditions. For the time of collection, This refers to the effective subcutaneous light intensity in the near-infrared band.

[0024] Step 3: Extract the facial rosy complexion index based on the CIELab color space. And the gloss index of body fluids Specifically, it includes the following steps: Based on the CIELab color space and combined with blood oxygen saturation data in the 940nm near-infrared band under imaging conditions, the traditional red-green axis is corrected as the quantification of the blood rosiness index. Furthermore, the body fluid luster index is retrieved using the 450nm blue light band. The formula for extracting the blood rosiness index is as follows: in, The index represents the degree of rosy complexion and health. These are the weighting coefficients. These are the red and green axis values ​​in the CIELab space. The blood oxygen saturation was obtained under 940nm near-infrared imaging conditions; the formula for extracting the body fluid luster index is: in, The gloss index of body fluids. These are the weighting coefficients. The brightness value for the CIELab space. The skin moisture content was obtained by inversion under imaging conditions of 450nm blue light band.

[0025] Step 4: Acquire striped projection images, perform 3D reconstruction, generate a two-layer 3D model, and overlay facial meridian lines. and assessment area Obtain the center line of the wrinkle and three-dimensional wrinkle morphology parameters Specifically, it includes the following steps.

[0026] Step 4.1: The camera and grating projector are combined to obtain a striped projection image. The striped projection image includes a frontal image, a left 45° facial image, a right 45° facial image, and a combined image with a custom tilt angle. The combined image with the custom tilt angle includes an image of a dense facial wrinkle area and multiple tilt angles of a facial acupuncture point area. The image of a dense facial wrinkle area is the main direction angle of the wrinkle dense area extracted by an edge detection algorithm.

[0027] Step 4.2: Perform 3D reconstruction on the striped projection image obtained in Step 4.1, and obtain the macroscopic facial contours using structured light. And extracting micro-wrinkles using phase deflection technique Blend to generate a two-layer 3D model ; The structured light acquisition of the macroscopic facial contour is calculated using the following formula: ,in, In the striped projection image Macroscopic contour depth of a pixel. For the fringe amplitude, The phase value of the structured light stripe. The stripe spacing; The phase deflection technique extracts micro-wrinkles using the following formula: ,in, for The depth of micro-wrinkles at the pixel level. These are the phase-to-depth conversion coefficients. This represents the phase deflection.

[0028] Step 4.3: Overlay facial meridian lines onto the double-layered 3D model obtained in Step 4.2. Obtain the center line of the wrinkle Identify the center line of wrinkles With meridian lines The intersection points are superimposed on the evaluation areas calibrated in step 1 onto the double-layered 3D model. Identify the depth of skin depressions and protrusions around acupoints, and record three-dimensional wrinkle morphology parameters. ; Identify the intersection of wrinkles and meridians using the following formula: and ,in The wrinkle center line is obtained by superimposing meridian lines on a 3D model. For meridian lines, The average depth within a 5mm radius around the intersection. The concave threshold is used; if the formula is satisfied, it is determined to be a valid intersection point. The following formula is used to identify the skin depressions and protrusions around acupoints and record the three-dimensional wrinkle morphology parameters. : ,in, The depression around the acupoint skin in the i-th evaluation region represents the three-dimensional wrinkle morphology parameter. To assess the average depth of the area, The minimum depth value for the core assessment area. For the assessment area.

[0029] Step 5: Guide the execution of the set facial expression sequence, capture the instantaneous changes in facial muscle movement, complete muscle dynamics acquisition, and obtain the angle between the muscle movement direction of each pixel and the meridian direction. Furthermore, a correlation coefficient between exercise intensity and acupoint distance was established within the assessment area. Specifically, it includes the following steps.

[0030] Step 5.1: Irradiate the face with a uniform near-infrared light source to avoid strong light stimulating facial muscles, and then guide and execute the set expression sequence actions. The set expression sequence includes: frowning and raising the forehead to activate the Du meridian and Bladder meridian in the forehead; pursing the lips and puffing out the cheeks to activate the Ren meridian and Stomach meridian around the mouth; lifting the side of the face and lifting the corners of the mouth diagonally upward to activate the Gallbladder meridian and Large Intestine meridian in the cheek; blinking and frowning to simultaneously activate the Liver meridian and Triple Energizer meridian around the eyes.

[0031] Step 5.2: Capture the instantaneous changes in muscle movement during the execution of the set facial expression sequence in Step 5.1 using a camera, record the motion trajectory of the evaluation area, obtain video frames, analyze the pixel motion between video frames, calculate the strain value and contraction velocity of the muscles in the evaluation area, map the strain field to the corresponding meridian lines, and map the contraction velocity to the corresponding meridian lines. Obtain the angle between the muscle movement direction of a pixel and the meridian direction. Furthermore, the correlation between exercise intensity and acupoint distance was established within the assessment area to complete muscle dynamics data acquisition. The analysis of pixel motion between video frames specifically involves calculating the strain values ​​of the muscle in the evaluation region of the pixel at any given time in the x and y axes based on the displacement of the pixel motion between video frames, and then calculating the strain field using the dense optical flow method. ; in , Let be the strain values ​​of the pixel at time t in the x-axis and y-axis directions. , Let x and y be the displacements of the pixel at time t along the x and y axes. , The pixel interval is [number].

[0032] The angle, strain value, and contraction speed of the muscle movement direction of each pixel are synchronously mapped to the meridian line. ,include: The angle between the direction of muscle movement at a pixel and the direction of the meridian is calculated using the following formula: ,in For pixels The angle between the direction of muscle movement and the direction of the meridian. The direction vector of muscle movement is mapped from the strain field to the corresponding meridian lines. get, The contraction speed is obtained by mapping the direction vector of the meridian to the corresponding meridian line; Then establish the correlation between exercise intensity and acupoint distance: , Let be the correlation coefficient between muscle movement intensity and acupoint distance in the i-th assessment region. For covariance, For variance, Let (x, y) be the distance from the (x, y) pixel to the center of the i-th acupoint. The intensity of muscle movement at that pixel, i.e. The amplitude.

[0033] Step 6, the blood flow velocity dynamic signal from Step 2 The blood and complexion index obtained in step 3 Surface gloss index The three-dimensional wrinkle morphology parameters obtained in step 4 Step 5 obtains the angle between the muscle movement direction of the pixel and the meridian direction. The correlation coefficient between muscle movement intensity and acupoint distance within the region was assessed. Construct a dataset for evaluating the efficacy of acupuncture. The acupuncture efficacy evaluation dataset was matched with thresholds set for TCM syndrome types, and a deep learning model was used to fuse multimodal data to generate evaluation results.

[0034] Preferably, when and , The threshold for the blood and qi rosiness index. Based on the blood flow velocity threshold, the matching assessment indicates a deficiency of Qi and Blood. when and and ,in , For blood flow velocity in different assessment areas, For the difference in maximum blood flow velocity, The threshold for blood flow difference. The number of effective intersections between wrinkles and meridians The threshold for the number of intersections. The depth threshold for wrinkles of the Qi stagnation and blood stasis type was used, and the matching assessment was classified as Qi stagnation and blood stasis type.

[0035] This is used to match and evaluate various certificate types.

[0036] The above description is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-angle optical imaging-based comprehensive evaluation method for facial acupuncture efficacy, characterized in that, Includes the following steps: Step 1, scan the human face, combine the facial meridian direction map, and take the area within 1.5 cm around the acupoint, meridian and facial acupuncture acupoint as the evaluation area ; Step 2, continuously shoot human face images, screen clear images and calibrate ambient light intensity in real time, then extract the blood flow velocity dynamic signal of subcutaneous capillary through polarization extinction model and spectral separation algorithm ; Step 3, Extracting the human face blood redness index based on CIELab color space and the body fluid luster index ; Step 4, collect the stripe projection image, perform three-dimensional reconstruction, generate a double-layer three-dimensional model, and superimpose facial meridian lines and evaluation area , obtain wrinkle center line and three-dimensional wrinkle morphology parameters ; Step 5, guide the execution of the set expression sequence action, capture the instantaneous change of human facial muscle movement, complete muscle power collection, and obtain the angle between the muscle movement direction of the pixel point and the meridian direction , and establish the correlation coefficient between the motion intensity and the distance of the acupoint in the evaluation area ; Step 6, blood flow velocity dynamic signal of step 2 , gas blood redness index obtained in step 3 , fluid glossiness index , three-dimensional wrinkle morphology parameters obtained in step 4 , the angle between the muscle movement direction of the pixel point obtained in step 5 and the meridian direction , the correlation coefficient of muscle movement intensity and distance from the acupoint in the evaluation area , construct an acupuncture efficacy evaluation dataset , match the acupuncture efficacy evaluation dataset with the TCM syndrome type setting threshold, use a deep learning model, fuse multi-modal data, and generate an evaluation result.

2. The comprehensive evaluation method for facial acupuncture efficacy according to claim 1, characterized in that, The following process is used for continuous image capture: A characteristic spectrum with three different polarization angles (0°, 45°, and 90°) is sequentially output through an LED array. This spectrum includes four types of light: 450nm blue light, 550nm green light, 650nm red light, and 940nm near-infrared light, generating 12 sets of imaging conditions. Under each imaging condition, the camera continuously captures images of the human face, resulting in continuous images. The gradient value of each pixel on the continuous image plane is used to select the clear images from the continuous images under each imaging condition. The ambient light intensity of the clear images under each imaging condition is recorded, and the ambient light intensity is calibrated in real time using a light sensor. The real-time calibrated clear images are then processed using a polarization extinction model and a spectral separation algorithm to highlight the oil reflection on the facial skin surface and extract the dynamic signal of blood flow velocity in subcutaneous capillaries.

3. The comprehensive evaluation method for facial acupuncture efficacy according to claim 2, characterized in that, The gradient value filtering is achieved through the following formula: ;in, , The Sobel operator is used to compute the images in... Pixel x axis, y Axial gradient value, For overall image sharpness, The sharpness threshold is used; if the formula is satisfied, the image is considered sharp. The real-time calibration is achieved through the following formula: ;in, The brightness values ​​of the original, clear image captured by the camera. For standard lighting intensity, This refers to the calibrated image brightness value.

4. The comprehensive evaluation method for facial acupuncture efficacy according to claim 2, characterized in that, The polarization extinction model: ,in, For clear images calibrated in real time, The intensity of mirror reflection on the skin surface. For extinction efficiency, To eliminate the effective light intensity in the subcutaneous tissue after reflection; The dynamic signal of blood flow velocity in subcutaneous capillaries was extracted using the following blood flow velocity model. ,in, This represents the dynamic signal of blood flow velocity in subcutaneous capillaries. For calibration coefficients, This represents the change in light intensity under near-infrared conditions. For the time of collection, This refers to the effective subcutaneous light intensity in the near-infrared band.

5. The comprehensive evaluation method for facial acupuncture efficacy according to claim 2, characterized in that, In step 3, based on the CIELab color space and combined with blood oxygen saturation data in the 940nm near-infrared band under imaging conditions, the traditional red-green axis is corrected as the quantification of the blood rosiness index. The body fluid luster index is then retrieved using the 450nm blue light band. The formula for extracting the blood rosiness index is as follows: in, The index represents the degree of rosy complexion and health. These are the weighting coefficients. These are the red and green axis values ​​in the CIELab space. The blood oxygen saturation was obtained under 940nm near-infrared imaging conditions; the formula for extracting the body fluid luster index is: in, The gloss index of body fluids. These are the weighting coefficients. The brightness value for the CIELab space. The skin moisture content was obtained by inversion under imaging conditions of 450nm blue light band.

6. The comprehensive evaluation method for facial acupuncture efficacy according to claim 5, characterized in that, In step 4, the camera and grating projector are combined to obtain a striped projection image. The striped projection image includes a frontal image, a left 45° facial image, a right 45° facial image, and a combined image with a custom tilt angle. The combined image with the custom tilt angle includes an image of a dense facial wrinkle area and multiple tilt angles of a facial acupuncture point area. The image of a dense facial wrinkle area is the main direction angle of the dense wrinkle area extracted by an edge detection algorithm.

7. The comprehensive evaluation method for facial acupuncture efficacy according to claim 6, characterized in that, In step 4, the striped projection image is reconstructed in three dimensions, and the macroscopic facial contour is obtained through structured light. And extracting micro-wrinkles using phase deflection technique Blend to generate a two-layer 3D model ; The structured light acquisition of the macroscopic facial contour is calculated using the following formula: ,in, In the striped projection image Macroscopic contour depth of a pixel. For the fringe amplitude, The phase value of the structured light stripe. The stripe spacing; The phase deflection technique extracts micro-wrinkles using the following formula: ,in, for The depth of micro-wrinkles at the pixel level. These are the phase-to-depth conversion coefficients. This represents the phase deflection.

8. The comprehensive evaluation method for facial acupuncture efficacy according to claim 7, characterized in that, In step 4, facial meridian lines are superimposed on the double-layered 3D model. Obtain the center line of the wrinkle Identify the center line of wrinkles With meridian lines The intersection points are superimposed on the evaluation areas calibrated in step 1 onto the double-layered 3D model. Identify the depth of skin depressions and protrusions around acupoints, and record three-dimensional wrinkle morphology parameters. ; Identify the intersection of wrinkles and meridians using the following formula: and ,in The wrinkle center line is obtained by superimposing meridian lines on a 3D model. For meridian lines, The average depth within a 5mm radius around the intersection. The concave threshold is used; if the formula is satisfied, it is determined to be a valid intersection point. The following formula is used to identify the skin depressions and protrusions around acupoints and record the three-dimensional wrinkle morphology parameters. : ,in, The depression around the acupoint skin in the i-th evaluation region represents the three-dimensional wrinkle morphology parameter. To assess the average depth of the area, The minimum depth value for the core assessment area. For the assessment area.

9. The comprehensive evaluation method for facial acupuncture efficacy according to claim 8, characterized in that, In step 5, the face is illuminated with a uniform near-infrared light source to avoid strong light stimulating the facial muscles. Then, the set expression sequence is guided and executed. The set expression sequence includes: frowning and raising the forehead to activate the Du meridian and Bladder meridian in the forehead; pursing the lips and puffing out the cheeks to activate the Ren meridian and Stomach meridian around the mouth; lifting the side of the face and lifting the corners of the mouth diagonally upward to activate the Gallbladder meridian and Large Intestine meridian in the cheek; blinking and frowning to simultaneously activate the Liver meridian and Triple Energizer meridian around the eyes. The camera captures instantaneous changes in muscle movement during the execution of a pre-defined sequence of facial expressions, records the motion trajectory of the evaluation area, obtains video frames, analyzes pixel motion between video frames, calculates the strain value and contraction velocity of the muscles in the evaluation area, and maps the strain field to the corresponding meridian lines, as well as the contraction velocity to the corresponding meridian lines. Obtain the angle between the muscle movement direction of a pixel and the meridian direction. Furthermore, the correlation between exercise intensity and acupoint distance was established within the assessment area to complete muscle dynamics data acquisition. The analysis of pixel motion between video frames specifically involves calculating the strain values ​​of the muscle in the evaluation region of the pixel at any given time in the x and y axes based on the displacement of the pixel motion between video frames, and then calculating the strain field using the dense optical flow method. ; in , Let be the strain values ​​of the pixel at time t in the x-axis and y-axis directions. , Let x and y be the displacements of the pixel at time t along the x and y axes. , The pixel interval is [number].

10. The comprehensive evaluation method for facial acupuncture efficacy according to claim 9, characterized in that, The angle, strain value, and contraction speed of the muscle movement direction of each pixel are synchronously mapped to the meridian line. ,include: The angle between the direction of muscle movement at a pixel and the direction of the meridian is calculated using the following formula: ,in For pixels The angle between the direction of muscle movement and the direction of the meridian. The direction vector of muscle movement is mapped from the strain field to the corresponding meridian lines. get, The contraction speed is obtained by mapping the direction vector of the meridian to the corresponding meridian line; Then establish the correlation between exercise intensity and acupoint distance: , Let be the correlation coefficient between muscle movement intensity and acupoint distance in the i-th assessment region. For covariance, For variance, Let (x, y) be the distance from the (x, y) pixel to the center of the i-th acupoint. The intensity of muscle movement at that pixel, i.e. The amplitude.

Images