Intelligent traditional Chinese medicine hand diagnosis method and diagnosis system thereof
The intelligent TCM hand diagnosis system, combining sensor and machine learning technologies, has achieved standardization and normalization of TCM hand diagnosis, solving the problem of reliance on experience in traditional TCM hand diagnosis and improving the accuracy and efficiency of diagnosis.
Patent Information
- Application Number
- CN202511068150.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-31
- Publication Date
- 2025-11-07
AI Technical Summary
Traditional Chinese medicine hand diagnosis relies on the doctor's personal experience and lacks standardization and normalization, resulting in inconsistent diagnostic results and making it difficult to meet the rapid diagnostic needs of modern life.
By combining sensor technology, image processing technology, and machine learning algorithms, an intelligent TCM hand diagnosis system is established, including information acquisition, data processing, and intelligent analysis modules. This system achieves standardized data processing and a unified diagnostic model. Data is acquired through high-definition camera equipment, pulse sensors, and temperature sensors, and diagnosis is performed using convolutional neural networks and multimodal feature fusion.
It has improved the accuracy and efficiency of TCM diagnosis, achieved standardization and normalization of diagnosis, facilitated communication and cooperation between different regions and institutions, and enhanced the consistency of diagnosis.
Smart Images
Figure CN120913770A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of traditional Chinese medicine diagnosis, in particular to an intelligent traditional Chinese medicine hand diagnosis method and a diagnosis system thereof. BACKGROUND
[0002] As a treasure of the Chinese nation, traditional Chinese medicine has a long history and is profound, and its unique theoretical system and diagnosis method has made important contributions to human health. Hand diagnosis, as an important part of traditional Chinese medicine diagnosis, has a long history. In the modern medical environment, traditional Chinese medicine hand diagnosis faces many challenges. The accuracy and reliability of traditional Chinese medicine hand diagnosis largely depend on the personal experience and professional level of doctors. The diagnosis results of different doctors may differ, and there is a lack of standardized and standardized diagnosis process, which limits the promotion of hand diagnosis in clinical application. In addition, with the acceleration of modern life rhythm and the increasing demand for health, the efficiency of traditional hand diagnosis cannot meet the demand of large-scale and rapid diagnosis. SUMMARY
[0003] In view of the deficiencies of the prior art, the present application provides an intelligent traditional Chinese medicine hand diagnosis method and a diagnosis system thereof. The system combines traditional Chinese medicine hand diagnosis with modern sensor technology, image processing technology, machine learning algorithm, etc., promotes the modernization development of traditional Chinese medicine diagnosis, and provides a new idea and method for the inheritance and innovation of traditional Chinese medicine. Through standardized processing of the diagnosis process and data, a unified diagnosis model and database are established, which helps to promote the standardization and standardization of traditional Chinese medicine diagnosis, improve the consistency and comparability of traditional Chinese medicine diagnosis, and facilitate the exchange and cooperation of traditional Chinese medicine between different regions and institutions.
[0004] To achieve the above purpose, the present application is realized by the following technical scheme: An intelligent traditional Chinese medicine hand diagnosis method and a diagnosis system thereof, an intelligent traditional Chinese medicine diagnosis system, characterized in that the intelligent traditional Chinese medicine diagnosis system comprises an information acquisition module, a data processing module, an intelligent analysis module and a diagnosis result generation module. The information acquisition module acquires various data related to hand diagnosis, providing a data basis for subsequent analysis and processing of the system, including high-definition camera equipment, pulse sensor, temperature sensor and information input device. The data processing module is used for realizing real-time acquisition, transmission and storage of the data collected by the acquisition equipment of the information acquisition module. The data processing module comprises data feature extraction function and data quantization analysis function. The intelligent analysis module analyzes the processed data, identifies different hand diagnosis features, and judges the association with diseases. The intelligent analysis module comprises multi-modal feature fusion function, symptom matching function and pathological symptom inference function. The diagnostic result generation module is used to display analysis results of the system and provide relevant application services.
[0005] Further, the high-definition camera of the information collection module captures images of the palm shape, skin color, palm lines, and finger nail shape of the hand, and converts the visual features into digital images instead of traditional manual observation. The pulse sensor collects the frequency, strength, and rhythm of the pulse of the hand, and converts traditional manual pulse diagnosis into a quantitative pulse electrical signal. The temperature sensor scans the skin temperature of different regions of the hand. The information input device is used to input the patient's age, gender, main symptoms, and medical history, thereby associating with the hand data.
[0006] Further, the data feature extraction module of the data processing module adopts a convolutional neural network to scan images through multiple convolution kernels to extract palm image features. The data image feature extraction formula is:
[0007]
[0008] f is the collected image, g is the convolution kernel, (i,j) is the output feature map coordinate, a and b are the convolution kernel size parameters, the hand background is removed through the convolution kernel, and then the palm area is located through image segmentation. A loss function cross-entropy is used for pixel classification: where N is the total number of pixels, C is the number of categories, y i,c is the true label, is the predicted probability. Further, the data quantization analysis function converts the time-domain pulse signal collected by the pulse sensor into a frequency domain. The calculation method of the time-domain pulse signal is: where x(n) is the time-domain pulse signal, X(k) is the frequency domain amplitude of the collected signal, and N is the number of sampling points. The instantaneous changes of the pulse waveform are captured through multi-resolution analysis: where a is the frequency scale parameter, b is the time shift parameter, and ψ is the wavelet basis function. Further, the multi-modal feature fusion of the intelligent analysis module inputs the image features, pulse parameters, temperature data, and patient complaint information into a comprehensive analysis model, and establishes a mapping relationship between "hand features-organ pathogenesis-patterns" through multi-modal fusion with the constructed traditional Chinese medicine hand diagnosis database. The image features, pulse features, and symptom data are fused through weighted fusion. The weighted fusion formula is: F=w1·F image +w2·F pulse+w3·F symptom w1+w2+w3=1 The weights are optimized by gradient descent: Where M is the number of samples, y i is the TCM syndrome type label.
[0009] Further, the symptom matching of the intelligent analysis module is trained by machine learning to determine a syndrome model, compare the hand feature combination of the typical syndrome in the database, and output the highest similarity symptom and probability value. The syndrome model adopts random forest to determine the syndrome by voting of multiple decision trees, and each tree is based on feature splitting to maximize information gain: Where S is the sample set, A is the feature, and V is the feature value is the entropy.
[0010] Further, the pathological symptom inference function of the intelligent analysis module combines the fusion mode and the symptom matching result to analyze the abnormal function of the corresponding viscera and bowels of the feature vector, and predict the potential health risks.
[0011] Further, the diagnostic result generation has a visual result generation function to present the diagnostic result in the form of text and images, mark the abnormal areas of the hand image, interpret the pulse waveform graph and parameters, and conclude the syndrome type, and generate personalized rehabilitation suggestions according to the conclusion.
[0012] Further, the hand diagnosis method comprises the following steps: Step 1: patient information input, the patient submits the basic information through the information input device to provide background reference for diagnosis; Step 2: hand image acquisition, pulse signal acquisition and temperature acquisition, use high-definition camera equipment to take pictures of the palm, fingers and nails from multiple angles, obtain the visual features of palm type, skin color, texture, lines, nail shape, and collect the frequency, strength, rhythm, waveform data of pulse signals through pulse sensors attached to the wrist, Guan and Chi, and use infrared temperature sensors or contact temperature probes to measure the temperature distribution of different areas of the hand; Step 3: data processing, pre-process and image segmentation to extract palm image features, use convolutional neural network to extract palm print complexity, line direction and skin pigmentation area features, and extract pulse signal through frequency conversion; Step 4: Multi-modal multi-source data fusion, integrate image features, pulse parameters, temperature data and patient's complaint information into a multi-dimensional feature vector and set to a unified data format, then input the feature vector into the pre-trained multi-modal fusion model, compare with the feature template of typical syndrome type in the traditional Chinese medicine hand diagnosis database, Calculate the matching probability of each syndrome type; Step 5: Pathological prediction and result generation, combine the fusion data to infer the function state of viscera, predict potential health problems, convert the quantitative features into traditional Chinese medicine dialectical terms, and generate visual text reports, And generate personalized conditioning suggestions according to the pathology.
[0013] The present application provides an intelligent traditional Chinese medicine hand diagnosis method and its diagnosis system. It has the following beneficial effects: 1. The present application provides an intelligent traditional Chinese medicine hand diagnosis method and its diagnosis system, which greatly improves the diagnosis efficiency and objective effect of diagnosis. The device replaces the traditional manual observation, smell, interrogation and pulse feeling, realizes multi-dimensional data collection, can collect hand image, pulse, temperature and multi-dimensional data such as patient's basic information and medical history at the same time, greatly shortens the collection time of diagnosis data and improves the diagnosis efficiency.
[0014] 2. The present application provides an intelligent traditional Chinese medicine hand diagnosis method and its diagnosis system. By setting data strengthening processing and analysis ability in the system, image feature extraction can be carried out at the same time, and sensor data can be quantitatively analyzed. A multi-modal fusion model is established. The image features, pulse parameters, temperature data and patient's complaint information are input into the comprehensive analysis model, combined with the traditional Chinese medicine hand diagnosis database constructed, the mapping relationship of "hand features - viscera pathogenesis - syndrome type" is established through multi-modal fusion, the syndrome type of the patient is comprehensively judged from multiple dimensions, and the accuracy of diagnosis is improved.
[0015] 3. The present application provides an intelligent traditional Chinese medicine hand diagnosis method and its diagnosis system. The system combines traditional Chinese medicine hand diagnosis with modern sensor technology, image processing technology, machine learning algorithm, etc., promotes the modernization development of traditional Chinese medicine diagnosis, provides new ideas and methods for the inheritance and innovation of traditional Chinese medicine, and establishes a unified diagnosis model and database through the standardized processing of diagnosis process and data, which is helpful to promote the standardization and standardization of traditional Chinese medicine diagnosis, improve the consistency and comparability of traditional Chinese medicine diagnosis, and facilitate the exchange and cooperation of traditional Chinese medicine between different regions and institutions. BRIEF DESCRIPTION OF DRAWINGS
[0016] Fig. 1 is a system structure diagram of the intelligent traditional Chinese medicine hand diagnosis system of the present application. DETAILED DESCRIPTION
[0017] With reference to the accompanying drawings: the technical solutions in the embodiments of the present application will be described clearly and completely, obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.
[0018] Embodiment 1: As shown in Figure 1, the embodiment of the present application provides an intelligent traditional Chinese medicine hand diagnosis system, the intelligent traditional Chinese medicine diagnosis system comprising an information acquisition module, a data processing module, an intelligent analysis module and a diagnosis result generation module; The information acquisition module acquires various data related to hand diagnosis, providing a data basis for subsequent analysis and processing of the system, including high-definition camera equipment, pulse sensors, temperature sensors and information input equipment; The data processing module is used to realize real-time acquisition, transmission and storage of the data collected by the acquisition equipment of the information acquisition module, the data processing module comprising data feature extraction function and data quantitative analysis function; The intelligent analysis module analyzes the processed data, identifies different hand diagnosis features, and judges the association with diseases, the intelligent analysis module comprising multi-modal feature fusion function, symptom matching function and pathological symptom inference function; The diagnosis result generation module is used to display the analysis results of the system and provide related application services.
[0019] Embodiment 2: The high-definition camera equipment of the information acquisition module captures images of the palm type, skin color, palm lines and finger nail shape of the hand, converts visual features into digital images instead of traditional manual observation, the pulse sensor collects the frequency, intensity and rhythm of the hand pulse, converts traditional manual pulse diagnosis into quantitative pulse electrical signals, the temperature sensor scans the skin temperature of different areas of the hand, and the information input equipment is used to input the patient's age, gender, chief complaint, medical history, thereby associating with the hand data; When judging respiratory system diseases, the change of the color of the palm is an important observation index, when the color of the palm is reddish or presents purple red, it is often possible that the lung is hot, the number of red blood cells in the blood increases, and the color of the palm is reddish The amount of redness in the palm is relatively increased, which is common in diseases such as wind-heat cold and lung heat cough. When the body is invaded by wind-heat evil, symptoms such as fever, cough, and yellow and thick sputum may occur, and the color of the palm may also be reddish. If the color of the palm is pale, it may indicate lung cold or lung qi deficiency. Lung cold is caused by invasion of cold evil into the lungs, leading to lung qi failing to descend, and blood circulation being sluggish, resulting in the color of the palm turning white. Lung qi deficiency is caused by the weakening of lung function, which cannot fully promote blood circulation, leading to insufficient blood supply to the hands, and the color of the palm turning pale. It is common in diseases such as wind-cold cold and lung qi deficiency cough. The shape of the fingers is also an important basis for the diagnosis of respiratory diseases. Swelling and deformation of the finger joints may indicate chronic respiratory diseases. In the long-term development of chronic respiratory diseases, the body is in a state of chronic hypoxia, which stimulates the proliferation of tissues around the finger joints, leading to swelling and deformation of the finger joints. In patients with COPD, due to long-term airway obstruction and lung function decline, the body is hypoxic, and the distal end of the finger gradually appears swelling and deformation. This sign is called clubbing. Thin fingertips may indicate lung qi deficiency, which can cause blood circulation to be insufficiently nourished, making the fingertips thin, It is common in patients with weak constitution and long-term illness. They often have symptoms such as shortness of breath, fatigue, and cough. When the temperature of the palm is too high and the sweating is too much, it may indicate that the stomach fire is too strong. Stomach fire is caused by the excessive heat in the stomach, leading to the excessive function of the stomach, and symptoms such as stomach pain, bad breath, and gum pain. If the palm is too dry, it may indicate that the stomach yin is insufficient. Stomach yin deficiency is caused by the deficiency of yin fluid in the stomach, which cannot nourish the stomach, leading to symptoms such as stomach pain, loss of appetite, dry mouth, and dry throat. Nail indentation and curvature may indicate chronic gastritis. Pale and unattractive nails may indicate anemia or malnutrition. Digestive system diseases often cause poor absorption of nutrients, leading to anemia or malnutrition. If the central part of the palm is indented, it may indicate that the spleen and stomach are weak. The central part of the palm corresponds to the reflex zone of the spleen and stomach. The indentation of this area indicates that the function of the spleen and stomach is relatively weak, and it cannot normally transport water and blood, leading to the loss of filling in the central part of the palm. Judging urinary system diseases from the length and shape of the little finger is an important point in hand diagnosis. Short little finger may reflect the insufficient development or function of the kidney, which in turn affects the normal function of the urinary system. If the little finger is curved and deformed, it may indicate a tendency of gynecological diseases, which is more common in women. The color and shape of the palm can also provide important information for the diagnosis of urinary system diseases. Dark and dull color of the palm may indicate kidney deficiency, which can cause blood circulation to be sluggish and lack of essence, resulting in the loss of luster and becoming dark. If the palm is swollen, it may indicate a tendency of nephritis.
[0020] Working principle: The steps of the hand diagnosis method are as follows: Step 1: Patient information entry. Patients submit basic information through the information input device to provide background reference for diagnosis. Step 2: Hand image acquisition, pulse signal acquisition, and temperature acquisition. High-definition camera equipment is used to take multi-angle pictures of the palm, fingers, and nails to obtain visual characteristics of palm shape, skin color, texture, lines, and nail shape. A pulse sensor is placed against the wrist at the cun, guan, and chi positions to collect pulse frequency, strength, rhythm, and waveform data. An infrared temperature sensor or contact temperature probe is used to measure the temperature distribution in different areas of the hand. Step 3: Data processing. Preprocess the image and segment the image to extract palm image features. Use a convolutional neural network to extract palm print complexity, line direction, and skin pigmentation features. Extract pulse signals through frequency conversion. Step 4: Multimodal and multi-source data fusion. Image features, pulse parameters, temperature data, and patient complaints are integrated into a multidimensional feature vector and set to a unified data format. Then, the feature vector is input into a pre-defined... The trained multimodal fusion model is compared with the feature templates of typical syndromes in the TCM hand diagnosis database to calculate the matching probability of each syndrome. Step 5: Pathological prediction and result generation. This involves combining fused data to infer the functional status of internal organs, predicting potential health problems, converting quantitative characteristics into TCM diagnostic terminology, and generating a visual report with charts and graphs. And personalized treatment suggestions are generated based on the pathology.
[0021] The following points should be noted in this article: 1. The accompanying drawings of the embodiments disclosed herein only relate to the structures involved in the embodiments disclosed herein; other structures can be referred to in general design.
[0022] 2. Where there is no conflict, the embodiments of this disclosure and the features in the embodiments can be combined with each other to obtain new embodiments.
[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
Claims
1. An intelligent traditional Chinese hand diagnosis system, characterized in that: The intelligent traditional Chinese medicine diagnosis system comprises an information collection module, a data processing module, an intelligent analysis module and a diagnosis result generation module; The information collection module collects various data related to hand diagnosis, provides a data basis for subsequent analysis and processing of the system, and comprises high-definition camera equipment, a pulse sensor, a temperature sensor and an information input device; The data processing module is used for realizing real-time collection, transmission and storage of data collected by the collection equipment of the information collection module, and comprises a data feature extraction function and a data quantitative analysis function; The intelligent analysis module analyzes the processed data, identifies different hand diagnosis features, and judges the correlation thereof with diseases, and comprises a multi-modal feature fusion function, a symptom matching function and a pathological symptom inference function; The diagnosis result generation module is used for displaying analysis results of the system and providing related application services.
2. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The high-definition camera equipment of the information collection module captures images of the palm shape, skin color, palm lines and finger nail shape of the hand, converts visual features into digital images to replace traditional manual observation, the pulse sensor collects the frequency, strength and rhythm of hand pulse, converts traditional manual pulse diagnosis into quantitative pulse electrical signals, the temperature sensor scans the skin temperature of different areas of the hand, and the information input device is used for inputting the age, gender, chief complaint, medical history of the patient, so as to associate with the hand data.
3. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The data feature extraction module of the data processing module adopts a convolutional neural network, scans images through multiple convolution kernels, extracts palm image features, and the data image feature extraction formula is: f is the collected image, g is the convolution kernel, (i,j) is the output feature map coordinate, a and b are the convolution kernel size parameters, the hand background is removed through the convolution kernel, the palm area is positioned through image segmentation, and the loss function cross entropy is used for pixel classification. where N is the total number of pixels, C is the number of classes, y i,c is the true label, is the predicted probability.
4. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The data quantitative analysis function converts the time domain pulse signal into the frequency domain, and the calculation method of the time domain pulse signal is: Wherein, x(n) is the time domain pulse signal, X(k) is the frequency domain amplitude of the collected signal, and N is the number of sampling points, and the instantaneous change of the pulse waveform is captured through multi-resolution analysis: Wherein, a is the frequency scale parameter, b is the time translation parameter, and ψ is the wavelet base function.
5. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The multi-modal feature fusion of the intelligent analysis module inputs image features, pulse parameters, temperature data and patient complaint information into a comprehensive analysis model, and establishes a mapping relationship of "hand feature-organ pathogenesis-pattern" through multi-modal fusion of the constructed traditional Chinese medicine hand diagnosis database, and the image features, pulse features and symptom data are fused through weighting, and the weighting fusion formula is: F = w1 · F image + w2 · F pulse + w3 · F symptom w1+w2+w3=1 The weight is optimized through gradient descent: where M is the number of samples, y i is the TCM syndrome type label.
6. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The symptom matching of the intelligent analysis module trains a syndrome differentiation model through machine learning, compares the typical hand feature combinations of the database with the database, outputs the symptoms and probability values with the highest similarity, and determines the syndrome type through voting of multiple decision trees by the random forest. Each tree is based on the maximum information gain of feature splitting: Wherein, S is a sample set, A is a feature, and V is a feature value number is an entropy.
7. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The pathological symptom inference function of the intelligent analysis module combines the fusion mode and the symptom matching result, analyzes the abnormal functions of the viscera and bowels corresponding to the feature vector, and predicts the potential health risks.
8. The intelligent traditional Chinese hand diagnosis system according to claim 1, characterized in that: The diagnostic result generation has a visual result generation function to present the diagnostic result in the form of graphics and text, mark the abnormal areas of the hand image, analyze the pulse waveform graph and parameter interpretation, and conclude the syndrome type. Individualized rehabilitation recommendations are generated according to the conclusion.
9. An intelligent traditional Chinese medicine hand diagnosis method, characterized in that: The method is realized by the intelligent traditional Chinese medicine hand diagnosis system of claim 1, and the hand diagnosis method comprises the following steps: Step 1: patient information input, the patient submits the basic information through the information input device to provide background reference for diagnosis; Step 2: hand image acquisition, pulse signal acquisition and temperature acquisition, high-definition camera equipment is used to take pictures of the palm, fingers and nails from multiple angles, to obtain the visual characteristics of the palm type, skin color, texture, lines, nail shape, and to collect the frequency, strength, rhythm, waveform data of the pulse signal through the pulse sensor attached to the wrist, Guan and Chi, and to measure the temperature distribution of different areas of the hand by using the infrared temperature sensor or the contact temperature probe; Step 3: data processing, image preprocessing and image segmentation are performed to extract the palm image features, convolutional neural network is used to extract the palm print complexity, line direction and skin pigmentation area features, and pulse signal is extracted through frequency conversion; Step 4: multi-modal multi-source data fusion, the image features, pulse parameters, temperature data and patient complaint information are integrated into a multi-dimensional feature vector and set to a unified data format, and then the feature vector is input into the pre-trained multi-modal fusion model to compare with the feature templates of the typical syndromes in the traditional Chinese medicine hand diagnosis database, the matching probability of each syndrome type is calculated; Step 5: pathological prediction and result generation, the function state of viscera and bowels is inferred by combining the fusion data, potential health problems are predicted, the quantitative features are converted into traditional Chinese medicine dialectical terms, and the results are generated into visual graphic reports, and individualized conditioning recommendations are generated according to the pathology.