Traditional Chinese medicine liver and gall disease detection method and system based on video computing technology

Abstract

The invention relates to the technical field of traditional Chinese medicine liver and gall disease detection, and provides a traditional Chinese medicine liver and gall disease detection method and system based on a video computing technology. Compensating the face video data through a nonlinear color compensation algorithm to obtain color compensation face video data; extracting a multi-modal pulse wave signal, and processing the multi-modal pulse wave signal through an enhancement and denoising fusion algorithm to obtain an enhanced and denoised pulse wave signal; motion artifact suppression is conducted on the enhanced and denoised pulse wave signals through a time-space domain motion artifact self-adaptive suppression algorithm, and artifact suppression pulse wave signals are obtained; calculating a heart rate variability parameter based on the artifact suppression pulse wave signal; the heart rate variability parameters are recognized through a traditional Chinese medicine liver and gall disease classification recognition algorithm, and a traditional Chinese medicine liver and gall disease detection result is obtained. The accuracy and efficiency of traditional Chinese medicine liver and gall disease detection are improved through heart rate variability parameter analysis guided by the traditional Chinese medicine syndrome theory and a disease classification recognition algorithm.

Description

Technical Field

[0001] This invention relates to the field of TCM liver and gallbladder disease detection technology, and in particular to a TCM liver and gallbladder disease detection method and system based on video computing technology. Background Technology

[0002] Diagnosis of hepatobiliary diseases in Traditional Chinese Medicine (TCM) is an important component of TCM clinical practice. Traditional TCM uses the four diagnostic methods of observation, auscultation and olfaction, inquiry, and palpation to identify the syndrome characteristics of hepatobiliary diseases. Among these, facial observation and pulse diagnosis are of significant value in identifying hepatobiliary diseases. TCM theory holds that the liver opens into the eyes, and that the heart and liver are interconnected in terms of blood vessels; therefore, changes in facial color and pulse characteristics can reflect the functional state of the liver and gallbladder. However, traditional manual diagnostic methods rely on the clinical experience and subjective judgment of physicians, resulting in problems such as inconsistent diagnostic standards, poor repeatability, and low efficiency. With the development of computer vision technology and biomedical signal processing technology, non-contact physiological parameter detection technology based on video images has gradually emerged, providing a new technical approach for objective diagnosis in TCM.

[0003] In existing technologies, video pulse wave detection mainly employs color change analysis and signal filtering methods to achieve basic heart rate detection. However, existing methods do not adequately consider the intrinsic connection mechanism between traditional Chinese medicine theory and modern signal processing technology, making it difficult to organically integrate objective physiological signal characteristics with actual TCM syndrome differentiation and treatment theories. They also neglect the special influence of bilirubin deposition on video signals in patients with hepatobiliary diseases and cannot accurately compensate for pathological color changes such as jaundice. Summary of the Invention

[0004] In view of this, the present invention proposes a method and system for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology. This method solves the problems of existing methods failing to adequately consider the intrinsic connection mechanism between traditional Chinese medicine theory and modern signal processing technology, making it difficult to organically integrate objective physiological signal characteristics with actual traditional Chinese medicine syndrome differentiation and treatment theory, ignoring the special influence of bilirubin deposition on video signals in patients with hepatobiliary diseases, and failing to accurately compensate for pathological color changes such as jaundice.

[0005] The technical solution of this invention is implemented as follows: On one hand, this invention provides a method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology, comprising the following steps: Obtain facial video data of the person to be tested; The facial video data is compensated using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data. Multimodal pulse wave signals are extracted from the color-compensated facial video data, and the multimodal pulse wave signals are processed by an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain enhanced and denoised pulse wave signals. The enhanced and denoised pulse wave signal is subjected to motion artifact suppression by a spatiotemporal domain motion artifact adaptive suppression algorithm based on the liver meridian direction, resulting in an artifact-suppressed pulse wave signal. Heart rate variability parameters are calculated based on the artifact suppression pulse wave signal. The heart rate variability parameters are identified using a traditional Chinese medicine (TCM) liver and gallbladder disease classification and recognition algorithm to obtain TCM liver and gallbladder disease detection results.

[0006] Based on the above technical solutions, preferably, the step of compensating the facial video data using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data includes: Extract the RGB color components of each pixel in the facial video data and convert them into hue, saturation, and luminance components in the HSV color space; Based on the correspondence between yellow and liver and gallbladder diseases in the five-color diagnosis theory of traditional Chinese medicine, the distribution characteristics of bilirubin in the facial area of ​​the person to be tested can be identified. A nonlinear spectral absorption compensation model was established based on the aforementioned bilirubin distribution characteristics. The nonlinear spectral absorption compensation model is used to perform pixel-by-pixel compensation processing on the HSV color space, and then converted back to the RGB color space to obtain the color-compensated facial video data.

[0007] Based on the above technical solutions, preferably, the step of establishing a nonlinear spectral absorption compensation model based on the bilirubin distribution characteristics includes: The estimated bilirubin concentration is calculated based on the hue component values ​​of different areas of the face of the person being tested. Establish an exponential decay mapping relationship between the estimated bilirubin concentration and the spectral absorption coefficient; A regional nonlinear compensation function is constructed based on the exponential decay mapping relationship; Different compensation function parameters were applied to the periorbital region, cheek region, and nasal alar region to obtain the nonlinear spectral absorption compensation model.

[0008] Based on the above technical solutions, preferably, the step of extracting multimodal pulse wave signals from the color-compensated facial video data and processing the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on traditional Chinese medicine's Qi and blood theory to obtain an enhanced and denoised pulse wave signal includes: Pulse wave signals are extracted from the red, green, and blue channels of the color-compensated facial video data to obtain multimodal pulse wave signals; According to the theory of Qi and Blood in Traditional Chinese Medicine, the multimodal pulse wave signal is decomposed into a gas modal signal and a blood modal signal. The gas modal signal corresponds to high-frequency microvascular pulsation, and the blood modal signal corresponds to low-frequency blood flow changes. Calculate the correlation coefficient between the gas modal signal and the blood modal signal; An adaptive weighted fusion algorithm is constructed based on the correlation coefficient to perform weighted fusion of the gas mode signal and the blood mode signal to obtain the enhanced and denoised pulse wave signal.

[0009] Based on the above technical solutions, preferably, the step of constructing an adaptive weight fusion algorithm based on the correlation coefficient includes: A verification model of the correlation between the heart and liver was established based on the theory of the five internal organs in traditional Chinese medicine. The cardiohepatic correlation verification model was used to cross-validate the gas modal signal and the blood modal signal to identify abnormal fluctuation components. Based on the amplitude and frequency characteristics of the abnormal fluctuation components, calculate the noise suppression factor; The correlation coefficient is combined with the noise suppression factor to dynamically adjust the fusion weights of the gas mode signal and the blood mode signal; The gas modal signal and the blood modal signal are adaptively weighted and fused using the dynamically adjusted fusion weights.

[0010] Based on the above technical solutions, preferably, the acquisition of facial video data of the person to be tested includes: The face of the person to be tested is detected, and the liver and gallbladder related regions of interest are determined according to the distribution theory of liver and gallbladder meridians in traditional Chinese medicine, and the coordinates of the facial regions of interest are obtained. The quality of video frames including the coordinates of the facial region of interest is evaluated, and valid video frames are selected according to preset quality evaluation criteria. The valid video frames are then combined to obtain the facial video data.

[0011] Based on the above technical solutions, preferably, the step of using a spatiotemporal motion artifact adaptive suppression algorithm based on the liver meridian pathway to suppress motion artifacts in the enhanced and denoised pulse wave signal, resulting in an artifact-suppressed pulse wave signal, includes: Based on the distribution and direction of the liver meridian in the face according to traditional Chinese medicine, the enhanced and denoised pulse wave signal is subjected to liver meridian-related motion pattern recognition to obtain motion pattern classification results. Based on the motion pattern classification results, a spatiotemporal adaptive filtering algorithm is used to differentially suppress different types of motion artifacts to obtain the artifact-suppressed pulse wave signal.

[0012] Based on the above technical solutions, preferably, the calculation of heart rate variability parameters based on the artifact-suppressed pulse wave signal includes: The heart rate peak point is detected from the artifact-suppressed pulse wave signal, and the heart rate rhythm of the artifact-suppressed pulse wave signal is analyzed according to the pulse diagnosis theory of traditional Chinese medicine to obtain heart rate rhythm characteristic parameters. Based on the theory of heart-liver correlation in traditional Chinese medicine, heart rate variability parameters related to liver and gallbladder diseases were calculated, and the heart rate variability parameters were obtained.

[0013] Based on the above technical solutions, preferably, the step of identifying the heart rate variability parameters using a traditional Chinese medicine liver and gallbladder disease classification and identification algorithm to obtain traditional Chinese medicine liver and gallbladder disease detection results includes: The heart rate variability parameters are input into a pre-constructed TCM syndrome feature mapping model, and feature space mapping is performed according to the TCM syndrome theory of liver and gallbladder diseases to obtain TCM syndrome feature vectors. Based on the TCM syndrome feature vector, the TCM liver and gallbladder disease classification and identification algorithm is used to determine the disease type and assess the severity, thereby obtaining the TCM liver and gallbladder disease detection results.

[0014] On the other hand, the present invention also provides a traditional Chinese medicine liver and gallbladder disease detection system based on video computing technology, the system comprising: The facial video acquisition module is used to acquire facial video data of the person being tested. The bilirubin spectrum color compensation module is used to compensate the facial video data using a nonlinear color compensation algorithm based on the bilirubin spectrum characteristics to obtain color-compensated facial video data. The multimodal pulse wave extraction and fusion module is used to extract multimodal pulse wave signals from the color-compensated facial video data, and process the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain an enhanced and denoised pulse wave signal. The liver meridian motion artifact suppression module is used to suppress motion artifacts in the enhanced and denoised pulse wave signal by using a spatiotemporal domain motion artifact adaptive suppression algorithm based on the direction of the liver meridian, so as to obtain an artifact-suppressed pulse wave signal. The heart rate variability parameter calculation module is used to calculate heart rate variability parameters based on the artifact-suppressed pulse wave signal. The Traditional Chinese Medicine (TCM) liver and gallbladder disease identification module is used to identify the heart rate variability parameters through a TCM liver and gallbladder disease classification and identification algorithm to obtain TCM liver and gallbladder disease detection results.

[0015] The present invention provides a method and system for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology, which has the following advantages over existing technologies: (1) By integrating traditional Chinese medicine theory with video computing methods, a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin is used to solve the interference of jaundice color change. A multimodal pulse wave enhancement and denoising fusion algorithm is constructed in combination with the theory of qi and blood in traditional Chinese medicine. A spatiotemporal motion artifact adaptive suppression mechanism is constructed based on the direction of the liver meridian. Through the heart rate variability parameter analysis and disease classification and identification algorithm guided by the theory of syndrome in traditional Chinese medicine, the accuracy and efficiency of traditional Chinese medicine liver and gallbladder disease detection are improved. (2) By integrating the theory of five-color diagnosis in traditional Chinese medicine with color space transformation technology, the HSV color space conversion algorithm is used to identify the distribution characteristics and concentration of facial bilirubin. A nonlinear spectral absorption compensation model is constructed by combining the exponential decay mapping relationship. Differentiated compensation function parameters are set for the periorbital, cheek and nasal regions according to the physiological characteristics of different anatomical regions of the face, which improves the reliability of pulse wave signal extraction. (3) By integrating the theory of Qi and Blood in traditional Chinese medicine with multi-channel signal processing methods, the RGB three-channel pulse wave signal extraction algorithm is used to decompose the Qi mode and blood mode signals. The heart-liver correlation verification model is constructed in combination with the theory of the five internal organs in traditional Chinese medicine to identify abnormal fluctuation components. The fusion weight parameters are dynamically adjusted according to the correlation coefficient and noise suppression factor, thereby improving the signal-to-noise ratio and feature expression ability of the pulse wave signal. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a traditional Chinese medicine liver and gallbladder disease detection method based on video computing technology according to the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This invention provides a method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology, comprising the following steps: Obtain facial video data of the person to be tested; The facial video data is compensated using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data. Multimodal pulse wave signals are extracted from the color-compensated facial video data, and the multimodal pulse wave signals are processed by an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain enhanced and denoised pulse wave signals. The enhanced and denoised pulse wave signal is subjected to motion artifact suppression by a spatiotemporal domain motion artifact adaptive suppression algorithm based on the liver meridian direction, resulting in an artifact-suppressed pulse wave signal. Heart rate variability parameters are calculated based on the artifact suppression pulse wave signal. The heart rate variability parameters are identified using a traditional Chinese medicine (TCM) liver and gallbladder disease classification and recognition algorithm to obtain TCM liver and gallbladder disease detection results.

[0020] Specifically, this embodiment integrates traditional Chinese medicine theory with video computing methods. It utilizes a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to solve the interference of jaundice color changes, constructs a multimodal pulse wave enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine, and builds a spatiotemporal motion artifact adaptive suppression mechanism based on the direction of the liver meridian. Through heart rate variability parameter analysis and disease classification and identification algorithms guided by the theory of syndrome differentiation in traditional Chinese medicine, it solves the problem of reliance on physicians' subjective experience and inconsistent standards in the diagnosis of hepatobiliary diseases in traditional Chinese medicine, thereby improving the accuracy and efficiency of detection of hepatobiliary diseases in traditional Chinese medicine.

[0021] The acquisition of facial video data of the person to be tested includes: The face of the person to be tested is detected, and the liver and gallbladder related regions of interest are determined according to the distribution theory of liver and gallbladder meridians in traditional Chinese medicine, and the coordinates of the facial regions of interest are obtained.

[0022] In one specific embodiment, based on the theory in traditional Chinese medicine that the liver opens into the eyes, the face of the person to be tested is divided into five candidate areas: the periorbital area, the temple area, the cheek area, the nasal wing area, and the jaw area. Based on the distribution path of the liver and gallbladder meridians on the face, different weight coefficients are set for the five candidate regions. The final hepatobiliary-related region of interest is determined based on the weighting coefficients, and the corresponding region coordinates are extracted as the coordinates of the facial region of interest.

[0023] The quality of video frames including the coordinates of the facial region of interest is evaluated, and valid video frames are selected according to preset quality evaluation criteria. The valid video frames are then combined to obtain the facial video data.

[0024] In one specific embodiment, the image sharpness index, illumination uniformity index, and motion blur index are calculated for each video frame; The image sharpness index, illumination uniformity index, and motion blur index are compared with their corresponding preset thresholds; When all three indicators meet the corresponding preset threshold requirements, the video frame is marked as a valid video frame. The facial video data is formed by combining consecutive valid video frames.

[0025] Specifically, this embodiment integrates traditional Chinese medicine's theory of liver and gallbladder meridians with computer vision technology. Utilizing the theory that the liver opens into the eyes, it guides facial region segmentation and weight coefficient setting. The optimal region of interest is determined by combining the distribution paths of the liver and gallbladder meridians, and effective video frames are selected based on multi-dimensional quality evaluation indicators such as image clarity, illumination uniformity, and motion blur. Through intelligent region positioning and quality control mechanisms, the problems of blind region selection and unstable data quality in traditional facial video acquisition are solved, improving the accuracy of pulse wave signal extraction.

[0026] The process of compensating the facial video data using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data includes: Extract the RGB color components of each pixel in the facial video data and convert them into hue, saturation, and luminance components in the HSV color space; Based on the correspondence between yellow and liver and gallbladder diseases in the five-color diagnosis theory of traditional Chinese medicine, the distribution characteristics of bilirubin in the facial area of ​​the person to be tested can be identified. A nonlinear spectral absorption compensation model was established based on the aforementioned bilirubin distribution characteristics. The nonlinear spectral absorption compensation model is used to perform pixel-by-pixel compensation processing on the HSV color space, and then converted back to the RGB color space to obtain the color-compensated facial video data.

[0027] The establishment of a nonlinear spectral absorption compensation model based on the bilirubin distribution characteristics includes: The estimated bilirubin concentration is calculated based on the hue component values ​​of different areas of the face of the person being tested. Establish an exponential decay mapping relationship between the estimated bilirubin concentration and the spectral absorption coefficient; A regional nonlinear compensation function is constructed based on the exponential decay mapping relationship; Different compensation function parameters were applied to the periorbital region, cheek region, and nasal alar region to obtain the nonlinear spectral absorption compensation model.

[0028] In one specific embodiment, the formula for calculating the estimated bilirubin concentration is: ; in, For pixels Estimated bilirubin concentration at the location; Bilirubin sensitivity coefficients for different facial regions; For pixels The hue component value at that location ranges from 0 to 360 degrees. For pixels The saturation component value at the location ranges from 0 to 1. This is a region-related saturation weight index; Brightness influencing factor; For pixels The brightness component value at that location ranges from 0 to 1. This is the offset of the baseline bilirubin concentration.

[0029] The calculation formula for the nonlinear spectral absorption compensation model is as follows: ; in, For pixels First The compensated color value of the channel; For pixels First The original color value of the channel; For the first The exponential decay coefficient of the channel; For pixels Estimated bilirubin concentration at the location; For the first The concentration response index of the channel; For the first The periodic compensation amplitude coefficient of the channel; For the first The periodic compensation frequency coefficient of the channel; For the first Periodic compensation phase shift of the channel.

[0030] Specifically, this embodiment integrates the theory of five-color diagnosis in traditional Chinese medicine with color space transformation technology. It uses the HSV color space conversion algorithm to identify facial bilirubin distribution characteristics and estimate concentration. It constructs a nonlinear spectral absorption compensation model by combining exponential decay mapping relationship. Based on the physiological characteristics of different anatomical regions of the face, it implements differentiated compensation function parameter settings for the periorbital, cheek, and nasal regions. Through a region-by-region pixel-by-pixel compensation processing mechanism, it solves the problem of facial color abnormalities in patients with liver and gallbladder diseases such as jaundice interfering with video signal quality, and improves the reliability of pulse wave signal extraction.

[0031] The step of extracting multimodal pulse wave signals from the color-compensated facial video data and processing the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on traditional Chinese medicine's Qi and blood theory to obtain an enhanced and denoised pulse wave signal includes: Pulse wave signals are extracted from the red, green, and blue channels of the color-compensated facial video data to obtain multimodal pulse wave signals; According to the theory of Qi and Blood in Traditional Chinese Medicine, the multimodal pulse wave signal is decomposed into a gas modal signal and a blood modal signal. The gas modal signal corresponds to high-frequency microvascular pulsation, and the blood modal signal corresponds to low-frequency blood flow changes. Calculate the correlation coefficient between the gas modal signal and the blood modal signal; An adaptive weighted fusion algorithm is constructed based on the correlation coefficient to perform weighted fusion of the gas mode signal and the blood mode signal to obtain the enhanced and denoised pulse wave signal.

[0032] The adaptive weight fusion algorithm constructed based on the correlation coefficient includes: A verification model of the correlation between the heart and liver was established based on the theory of the five internal organs in traditional Chinese medicine. The cardiohepatic correlation verification model was used to cross-validate the gas modal signal and the blood modal signal to identify abnormal fluctuation components. Based on the amplitude and frequency characteristics of the abnormal fluctuation components, calculate the noise suppression factor; The correlation coefficient is combined with the noise suppression factor to dynamically adjust the fusion weights of the gas mode signal and the blood mode signal; The gas modal signal and the blood modal signal are adaptively weighted and fused using the dynamically adjusted fusion weights.

[0033] In one specific embodiment, the correlation coefficient between the gas modal signal and the blood modal signal is calculated using the following formula: ; in, For a moment The correlation coefficient of the blood and qi modality; Total number of channels; For the first The weighting coefficients of the five internal organs in Traditional Chinese Medicine for the channels; For the first The gas mode signal value of the channel; For the first The time-domain average value of the channel air mode signal; For a moment The Blood modal signal values ​​of the channel; For the first The time-domain average value of the channel blood modal signal; For the first The heart-liver correlation delay parameter of the channel.

[0034] The formula for adaptive weighted fusion of gas mode signals and blood mode signals is as follows: ; ; ; in, For a moment The fused pulse wave signal; Number of signal channels; For the first Dynamic weighting of channel gas mode signals; For the first Dynamic weighting of channel blood modal signals; For the first The correlation coefficient between Qi and blood in the channel; For the first Abnormal fluctuation suppression parameters for channel gas mode signals; For the first Parameters for suppressing abnormal fluctuations in channel blood modal signals; For the first The abnormal fluctuation amplitude coefficient of the channel; For the first Noise sensitivity parameters of channel gas mode signals; For the first Noise sensitivity parameters of channel blood modal signals; For the first The noise suppression factor of the channel; The weighting normalization constant for the gas mode in traditional Chinese medicine theory; The weight normalization constant of the blood modality in traditional Chinese medicine theory is denoted as .

[0035] Specifically, this embodiment integrates the theory of Qi and blood in traditional Chinese medicine with multi-channel signal processing methods. It uses an RGB three-channel pulse wave signal extraction algorithm to decompose the Qi and blood modes. It combines the theory of the five internal organs in traditional Chinese medicine to construct a heart-liver correlation verification model to identify abnormal fluctuation components. It dynamically adjusts the fusion weight parameters according to the correlation coefficient and noise suppression factor. Through an adaptive weighted fusion mechanism, it solves the problems of traditional single-channel pulse wave signals being susceptible to noise interference and unstable signal quality, thereby improving the signal-to-noise ratio and feature expression capability of the pulse wave signal.

[0036] The enhanced and denoised pulse wave signal is subjected to motion artifact suppression using a spatiotemporal domain motion artifact adaptive suppression algorithm based on the liver meridian pathway, resulting in an artifact-suppressed pulse wave signal, including: Based on the distribution and direction of the liver meridian in the face according to traditional Chinese medicine, the enhanced and denoised pulse wave signal is subjected to liver meridian-related motion pattern recognition to obtain motion pattern classification results.

[0037] In one specific embodiment, a facial liver meridian distribution map is constructed based on the theory of liver meridians in traditional Chinese medicine, dividing the face into the main pathway area and the secondary pathway area of ​​the liver meridian. Monitor the amplitude changes of pulse wave signals in the main and secondary pathways of the liver meridian; When a frowning movement related to liver qi stagnation is detected, it is classified as an emotional movement pattern; when a large head rotation is detected, it is classified as a holistic movement pattern. The classification result of the motion pattern is determined based on the amplitude of motion and the area of ​​influence.

[0038] Based on the motion pattern classification results, a spatiotemporal adaptive filtering algorithm is used to differentially suppress different types of motion artifacts to obtain the artifact-suppressed pulse wave signal.

[0039] In one specific embodiment, for the emotional movement pattern, a method combining high-pass filtering and liver meridian weight adjustment is used to suppress small-amplitude tremors; For the overall motion pattern, a method combining motion trajectory prediction and time-domain compensation is used for large-scale motion correction; Based on the degree of influence of different motion patterns on the liver and gallbladder-related facial areas, the filtering parameters and compensation intensity are dynamically adjusted. The processed signals from each region are weighted and fused to obtain the artifact-suppressed pulse wave signal.

[0040] Specifically, this embodiment integrates traditional Chinese medicine (TCM) liver meridian theory with spatiotemporal signal processing technology. It utilizes a facial distribution map of the liver meridian to construct the division of primary and secondary pathways. Combined with TCM pathological features such as liver qi stagnation, it intelligently identifies emotional and holistic movement patterns. Differentiated high-pass filtering, motion trajectory prediction, and temporal compensation algorithms are employed based on the influencing mechanisms of different movement patterns. Through adaptive parameter adjustment and weighted fusion mechanisms, it addresses the problems of traditional motion artifact suppression methods neglecting TCM meridian characteristics and lacking targeted processing, thereby improving the stability of pulse wave signals in complex motion environments.

[0041] The calculation of heart rate variability parameters based on the artifact-suppressed pulse wave signal includes: The heart rate peak point is detected from the artifact-suppressed pulse wave signal, and the heart rate rhythm of the artifact-suppressed pulse wave signal is analyzed according to the pulse theory of traditional Chinese medicine to obtain the characteristic parameters of heart rate rhythm.

[0042] In one specific embodiment, an adaptive threshold algorithm is used to identify the heart rate peak point in the artifact-suppressed pulse wave signal; Calculate the time interval between adjacent heart rate peaks to obtain the RR interval sequence; Based on the pulse characteristics of rapid pulse, slow pulse, knotted pulse, and intermittent pulse in traditional Chinese medicine pulse theory, pulse pattern recognition is performed on the RR interval sequence. The heart rate rhythm characteristic parameters are obtained by extracting pulse characteristics such as pulse intensity changes, rhythm regularity, and pulsatility from traditional Chinese medicine pulse characteristics.

[0043] Based on the theory of heart-liver correlation in traditional Chinese medicine, heart rate variability parameters related to liver and gallbladder diseases were calculated, and the heart rate variability parameters were obtained.

[0044] In one specific embodiment, a correlation model between heart rate variability and liver and gallbladder function is established based on the theory of the interconnectedness of the heart and liver blood vessels in traditional Chinese medicine. Calculate the time-domain heart rate variability parameters, including the standard deviation of the RR interval and the root mean square of the difference between adjacent RR intervals; Calculate the frequency domain heart rate variability parameters, including low-frequency power, high-frequency power, and the ratio of low to high frequency power; By combining the syndrome characteristics of liver qi stagnation, liver blood deficiency, and damp-heat in liver and gallbladder diseases in traditional Chinese medicine, syndrome-weighted processing of time-domain heart rate variability parameters and frequency-domain heart rate variability parameters is performed to obtain heart rate variability parameters specific to liver and gallbladder diseases.

[0045] Specifically, this embodiment integrates traditional Chinese medicine pulse diagnosis theory with heart rate variability analysis technology. It utilizes an adaptive threshold algorithm for heart rate peak detection and RR interval sequence extraction, and combines traditional pulse characteristics such as rapid pulse, slow pulse, knotted pulse, and intermittent pulse for rhythm pattern recognition. Furthermore, it establishes a correlation model between heart rate variability and liver / gallbladder function based on the theory of heart-liver blood circulation. By applying syndrome-weighted processing to time-domain and frequency-domain parameters using syndrome characteristics such as liver qi stagnation, liver blood deficiency, and damp-heat in the liver and gallbladder, it addresses the shortcomings of traditional heart rate variability analysis, which lacks guidance from traditional Chinese medicine theory and has insufficient disease specificity. This improves the sensitivity and specificity of heart rate variability parameters for liver and gallbladder diseases.

[0046] The process of identifying the heart rate variability parameters using a traditional Chinese medicine (TCM) liver and gallbladder disease classification and recognition algorithm to obtain TCM liver and gallbladder disease detection results includes: The heart rate variability parameters are input into a pre-constructed TCM syndrome feature mapping model, and feature space mapping is performed according to the TCM syndrome theory of liver and gallbladder diseases to obtain TCM syndrome feature vectors.

[0047] In one specific embodiment, a multidimensional syndrome feature space is constructed based on the four main syndrome types of liver and gallbladder diseases in traditional Chinese medicine: liver qi stagnation, liver blood stasis, liver and gallbladder damp-heat, and liver and kidney yin deficiency. The heart rate variability parameters are weighted according to the pulse characteristics of different syndromes, with the following weights: liver qi stagnation corresponds to the wiry pulse, liver blood stasis corresponds to the hesitant pulse, liver and gallbladder damp heat corresponds to the slippery and rapid pulse, and liver and kidney yin deficiency corresponds to the thin and weak pulse. The weighted parameters are converted into the TCM syndrome feature vector through a nonlinear feature mapping function.

[0048] Based on the TCM syndrome feature vector, the TCM liver and gallbladder disease classification and identification algorithm is used to determine the disease type and assess the severity, thereby obtaining the TCM liver and gallbladder disease detection results.

[0049] In one specific embodiment, a multi-level classifier based on the TCM syndrome differentiation and treatment theory is established. The first-level classifier is used to determine whether there is a liver and gallbladder disease, and the second-level classifier is used to identify the specific TCM syndrome type. Based on the numerical distribution of the TCM syndrome feature vectors, the similarity score of each syndrome type is calculated; The severity of diseases is quantitatively rated by combining the theories of deficiency and excess, cold and heat in traditional Chinese medicine. Based on the syndrome type identification results and severity rating, the TCM liver and gallbladder disease detection results are generated, including disease type, syndrome type and severity level.

[0050] Specifically, this embodiment integrates traditional Chinese medicine (TCM) syndrome differentiation and treatment theory with pattern recognition. It constructs a multi-dimensional feature space using four major syndrome types: liver qi stagnation, liver blood stasis, liver and gallbladder damp-heat, and liver and kidney yin deficiency. It then performs nonlinear mapping transformation by combining pulse characteristic weights such as wiry pulse, hesitant pulse, slippery and rapid pulse, and thin and weak pulse. A multi-level classifier architecture is employed to achieve disease presence discrimination and specific syndrome type identification. Through a severity quantification rating mechanism guided by the theories of deficiency / excess and cold / heat, it addresses the lack of TCM theoretical content and syndrome differentiation and treatment characteristics in traditional disease detection methods. This achieves intelligent identification of liver and gallbladder diseases that aligns with TCM diagnostic thinking, improving the consistency of diagnostic results with TCM theory.

[0051] This invention also provides a traditional Chinese medicine liver and gallbladder disease detection system based on video computing technology, the system comprising: The facial video acquisition module is used to acquire facial video data of the person being tested. The bilirubin spectrum color compensation module is used to compensate the facial video data using a nonlinear color compensation algorithm based on the bilirubin spectrum characteristics to obtain color-compensated facial video data. The multimodal pulse wave extraction and fusion module is used to extract multimodal pulse wave signals from the color-compensated facial video data, and process the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain an enhanced and denoised pulse wave signal. The liver meridian motion artifact suppression module is used to suppress motion artifacts in the enhanced and denoised pulse wave signal by using a spatiotemporal domain motion artifact adaptive suppression algorithm based on the direction of the liver meridian, so as to obtain an artifact-suppressed pulse wave signal. The heart rate variability parameter calculation module is used to calculate heart rate variability parameters based on the artifact-suppressed pulse wave signal. The Traditional Chinese Medicine (TCM) liver and gallbladder disease identification module is used to identify the heart rate variability parameters through a TCM liver and gallbladder disease classification and identification algorithm to obtain TCM liver and gallbladder disease detection results.

[0052] Specifically, this embodiment presents a TCM hepatobiliary disease detection system based on video computing technology. By constructing a modular intelligent detection system that integrates traditional TCM theory with video computing technology, it utilizes six core functional modules: bilirubin spectral color compensation, multimodal pulse wave fusion guided by TCM Qi and blood theory, motion artifact suppression of liver meridian pathways, heart rate variability parameter calculation, and TCM syndrome classification and recognition. This achieves fully automated processing from facial video acquisition to the output of hepatobiliary disease diagnosis results. Through inter-module collaboration and algorithm optimization and integration, it solves the problems of reliance on subjective experience, inconsistent standards, and low efficiency in traditional TCM hepatobiliary disease diagnosis, thereby improving the objectivity of diagnosis.

[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology, characterized in that, Includes the following steps: Obtain facial video data of the person to be tested; The facial video data is compensated using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data. Multimodal pulse wave signals are extracted from the color-compensated facial video data, and the multimodal pulse wave signals are processed by an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain enhanced and denoised pulse wave signals. The enhanced and denoised pulse wave signal is subjected to motion artifact suppression by a spatiotemporal domain motion artifact adaptive suppression algorithm based on the liver meridian direction, resulting in an artifact-suppressed pulse wave signal. Heart rate variability parameters are calculated based on the artifact suppression pulse wave signal. The heart rate variability parameters are identified using a traditional Chinese medicine (TCM) liver and gallbladder disease classification and recognition algorithm to obtain TCM liver and gallbladder disease detection results.

2. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The process of compensating the facial video data using a nonlinear color compensation algorithm based on the spectral characteristics of bilirubin to obtain color-compensated facial video data includes: Extract the RGB color components of each pixel in the facial video data and convert them into hue, saturation, and luminance components in the HSV color space; Based on the correspondence between yellow and liver and gallbladder diseases in the five-color diagnosis theory of traditional Chinese medicine, the distribution characteristics of bilirubin in the facial area of ​​the person to be tested can be identified. A nonlinear spectral absorption compensation model was established based on the aforementioned bilirubin distribution characteristics. The nonlinear spectral absorption compensation model is used to perform pixel-by-pixel compensation processing on the HSV color space, and then converted back to the RGB color space to obtain the color-compensated facial video data.

3. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 2, characterized in that, The establishment of a nonlinear spectral absorption compensation model based on the bilirubin distribution characteristics includes: The estimated bilirubin concentration is calculated based on the hue component values ​​of different areas of the face of the person being tested. Establish an exponential decay mapping relationship between the estimated bilirubin concentration and the spectral absorption coefficient; A regional nonlinear compensation function is constructed based on the exponential decay mapping relationship; Different compensation function parameters were applied to the periorbital region, cheek region, and nasal alar region to obtain the nonlinear spectral absorption compensation model.

4. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The step of extracting multimodal pulse wave signals from the color-compensated facial video data and processing the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on traditional Chinese medicine's Qi and blood theory to obtain an enhanced and denoised pulse wave signal includes: Pulse wave signals are extracted from the red, green, and blue channels of the color-compensated facial video data to obtain multimodal pulse wave signals; According to the theory of Qi and Blood in Traditional Chinese Medicine, the multimodal pulse wave signal is decomposed into a gas modal signal and a blood modal signal. The gas modal signal corresponds to high-frequency microvascular pulsation, and the blood modal signal corresponds to low-frequency blood flow changes. Calculate the correlation coefficient between the gas modal signal and the blood modal signal; An adaptive weighted fusion algorithm is constructed based on the correlation coefficient to perform weighted fusion of the gas mode signal and the blood mode signal to obtain the enhanced and denoised pulse wave signal.

5. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 4, characterized in that, The adaptive weight fusion algorithm constructed based on the correlation coefficient includes: A verification model of the correlation between the heart and liver was established based on the theory of the five internal organs in traditional Chinese medicine. The cardiohepatic correlation verification model was used to cross-validate the gas modal signal and the blood modal signal to identify abnormal fluctuation components. Based on the amplitude and frequency characteristics of the abnormal fluctuation components, calculate the noise suppression factor; The correlation coefficient is combined with the noise suppression factor to dynamically adjust the fusion weights of the gas mode signal and the blood mode signal; The gas modal signal and the blood modal signal are adaptively weighted and fused using the dynamically adjusted fusion weights.

6. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The acquisition of facial video data of the person to be tested includes: The face of the person to be tested is detected, and the liver and gallbladder related regions of interest are determined according to the distribution theory of liver and gallbladder meridians in traditional Chinese medicine, and the coordinates of the facial regions of interest are obtained. The quality of video frames including the coordinates of the facial region of interest is evaluated, and valid video frames are selected according to preset quality evaluation criteria. The valid video frames are then combined to obtain the facial video data.

7. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The enhanced and denoised pulse wave signal is subjected to motion artifact suppression using a spatiotemporal domain motion artifact adaptive suppression algorithm based on the liver meridian pathway, resulting in an artifact-suppressed pulse wave signal, including: Based on the distribution and direction of the liver meridian in the face according to traditional Chinese medicine, the enhanced and denoised pulse wave signal is subjected to liver meridian-related motion pattern recognition to obtain motion pattern classification results. Based on the motion pattern classification results, a spatiotemporal adaptive filtering algorithm is used to differentially suppress different types of motion artifacts to obtain the artifact-suppressed pulse wave signal.

8. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The calculation of heart rate variability parameters based on the artifact-suppressed pulse wave signal includes: The heart rate peak point is detected from the artifact-suppressed pulse wave signal, and the heart rate rhythm of the artifact-suppressed pulse wave signal is analyzed according to the pulse theory of traditional Chinese medicine to obtain heart rate rhythm characteristic parameters. Based on the theory of heart-liver correlation in traditional Chinese medicine, heart rate variability parameters related to liver and gallbladder diseases were calculated, and the heart rate variability parameters were obtained.

9. The method for detecting hepatobiliary diseases in traditional Chinese medicine based on video computing technology as described in claim 1, characterized in that, The process of identifying the heart rate variability parameters using a traditional Chinese medicine (TCM) liver and gallbladder disease classification and recognition algorithm to obtain TCM liver and gallbladder disease detection results includes: The heart rate variability parameters are input into a pre-constructed TCM syndrome feature mapping model, and feature space mapping is performed according to the TCM syndrome theory of liver and gallbladder diseases to obtain TCM syndrome feature vectors. Based on the TCM syndrome feature vector, the disease type is identified and the severity is assessed using a TCM liver and gallbladder disease classification and recognition algorithm to obtain the TCM liver and gallbladder disease detection results.

10. A traditional Chinese medicine (TCM) liver and gallbladder disease detection system based on video computing technology, used to execute a TCM liver and gallbladder disease detection method based on video computing technology as described in any one of claims 1-9, characterized in that, The system includes: The facial video acquisition module is used to acquire facial video data of the person being tested. The bilirubin spectrum color compensation module is used to compensate the facial video data using a nonlinear color compensation algorithm based on the bilirubin spectrum characteristics to obtain color-compensated facial video data. The multimodal pulse wave extraction and fusion module is used to extract multimodal pulse wave signals from the color-compensated facial video data, and process the multimodal pulse wave signals using an enhancement and denoising fusion algorithm based on the theory of Qi and blood in traditional Chinese medicine to obtain an enhanced and denoised pulse wave signal. The liver meridian motion artifact suppression module is used to suppress motion artifacts in the enhanced and denoised pulse wave signal by using a spatiotemporal domain motion artifact adaptive suppression algorithm based on the direction of the liver meridian, so as to obtain an artifact-suppressed pulse wave signal. The heart rate variability parameter calculation module is used to calculate heart rate variability parameters based on the artifact-suppressed pulse wave signal. The Traditional Chinese Medicine (TCM) liver and gallbladder disease identification module is used to identify the heart rate variability parameters through a TCM liver and gallbladder disease classification and identification algorithm to obtain TCM liver and gallbladder disease detection results.

Images