Method for extracting palmprint features based on digital image processing algorithm, predicting physical symptoms and judging correspondence between medical physiques

Abstract

The invention relates to the field of palmprint feature recognition, and discloses and aims to provide a method for preventing possible diseases and possible traditional Chinese medicine physiques by comparing extracted palmprint features with an established palmprint feature expression symptom database and analyzing the extracted palmprint features and the established palmprint feature expression symptom database. The method aims to predict possible diseases by researching palm print features and analyze the physique of the traditional Chinese medicine. According to the invention, a digital image processing algorithm is utilized to extract key features from the palm image, and the key features are combined with the traditional Chinese medicine constitution theory to achieve the purpose of disease prevention. In addition, the invention also discusses the application value of the palmprint characteristics in traditional Chinese medicine diagnosis, and opens up a new path for the combination of traditional medicine and modern science and technology.

Description

Technical field:

[0001] The present invention relates to the field of palm print feature recognition. The present invention is disclosed to provide a method for preventing possible diseases and possible TCM constitutions by comparing and analyzing extracted palm print features with an established database of palm print feature symptoms. The method aims to predict possible diseases by studying palm print features and analyze what TCM constitution a person may have. The present invention utilizes a digital image processing algorithm to extract key features from palm images and combines them with TCM constitution theory to achieve the purpose of preventing diseases. In addition, the present invention also explores the application value of palm print features in TCM diagnosis, opening up a new path for the combination of traditional medicine and modern technology. Background technology:

[0002] Data sources: existing standards of the China Association of Traditional Chinese Medicine; academic resource platforms such as academic conference proceedings and academic papers.

[0003] Related technologies: Color space conversion, threshold segmentation, geometric calculation, peak detection, feature point matching, etc. are all mature technical methods.

[0004] Reference materials: "Classification and Determination of Constitution in Traditional Chinese Medicine", "Preliminary Study on Palm Color of 79 Healthy People", "Observation of Hand Color in Diseased People", etc.

[0005] The key is how to match the possible symptoms with the analysis of palm print characteristics and analyze the possible TCM constitution. Summary of the invention:

[0006] At present, although palmprint feature extraction has certain potential, it still faces many challenges and shortcomings in practical applications. It is mainly reflected in the difficulty of data collection, the great impact of image quality, the limitations of feature extraction algorithms, large individual differences, and the challenges of large-scale application. First, high-quality palmprint image acquisition requires professional equipment, such as high-resolution cameras and stable light sources, which increases cost and complexity. Secondly, changes in lighting conditions will affect the quality of palmprint images, resulting in a decrease in the accuracy of feature extraction. Then, palmprint features are rich, but not all features are useful for recognition, and how to select effective features is a challenge. In addition, large-scale applications require the management of a large amount of palmprint data, and how to efficiently store and retrieve this data is a challenge. Finally, there is no unified mapping table to correspond palmprint features with predicted symptoms that may occur, and determine the possible type of TCM constitution.

[0007] This method extracts palm print features from images, providing a non-invasive way of health assessment, avoiding blood drawing or other forms of physical intervention in traditional diagnostic methods. Through instant image analysis, results can be obtained quickly, giving users instant feedback, which helps to adjust their living habits or seek medical help in time; users only need to provide two photos, which is simple to operate and easy to use, especially suitable for the elderly or people who are not familiar with complex technology. With the popularity of smartphones, this function can be integrated into various applications, making it convenient for the public to conduct self-tests at any time. At present, this system has achieved consistent performance on the web and mini-program sides, and APP is being implemented on Android, iOS, and Hongmeng systems. Although this method cannot replace professional medical diagnosis, it can be used as an auxiliary tool to help doctors obtain preliminary health information of patients more quickly. In addition, providing personalized health advice based on individual differences will help promote the development of personalized health management. By continuously recording changes in users' palm print features, it is also convenient to track health trends over a long period of time and detect abnormalities in a timely manner.

[0008] This system first extracts the left hand palm print features, then extracts the right hand palm print features, and finally returns the extraction results to the front end; (Note: The palm print feature extraction method for both hands is the same)

[0009] The system includes: (1) Hand and background separation module: The module receives the clear hand image uploaded by the front end and separates the hand outline from the background based on the skin color. (2) Effective area determination module: This module aims to determine the palm area through feature points in order to partition the palm. The geometric features of the palmprint are extracted through the peaks and troughs of the distance between the palm contour point and the wrist center point, and then other feature points on the palm are calculated based on the trough index and the wrist point to locate the key feature points of the palm. (3) Palm nine-area division module: The palm nine-area division is performed according to the feature points in (3) (for details, see Figure 3 , Figure 4 ). In order to extract the color and texture features of the palm print later. (4) Palm print color judgment module: Calculate the HSV mean of each area image of the nine areas of the palm, determine and extract the color features of the area; match the color features of each area with possible symptoms and TCM constitution according to Table 1, and return all results to the front end. (5) Palm print texture judgment module: This module is designed to detect the texture features of each of the nine areas of the palm. First, the image is preprocessed and converted into a grayscale image; then edge detection is performed after adaptive histogram equalization; finally, the polygonal approximation method is used to determine what kind of texture feature it belongs to based on the edge detection result. According to Table 1, the color features of a single area are matched with possible symptoms and TCM constitutions, and all results are fed back to the front end. For details, see the palmprint feature extraction system architecture Figure 1 . Palmprint feature extraction process diagram Figure 2 . The nine areas of the left palm Figure 3 . The nine areas of the right palm Figure 4 . Table 1: The characteristics of the palm lines in the nine zones, possible symptoms and TCM constitutions. Palm print characteristics, possible symptoms and TCM constitution correspondence table of the nine palm zones Table 1 Specific implementation method:

[0010] Step 1: Integrate the entire palmprint detection system into the company's health management system.

[0011] Step 2: Receive the clear palm image uploaded by the front end.

[0012] Step 3: Extract the palm area based on skin color.

[0013] Step 4: Extract the center point of the wrist, calculate the distance between all the contour points of the palm and the center point, and find the peaks and troughs.

[0014] Step 5: Calculate other feature points on the palm based on the trough index and wrist point.

[0015] Step 6: Divide the palm into eight trigrams zones according to all the feature points on the palm.

[0016] Step 7: Perform color recognition and texture recognition on each partition to obtain color features and texture features respectively.

[0017] Step 8: Feedback all recognition results to the system front end. Description of the drawings:

[0018] Palmprint feature extraction system architecture Figure 1 .

[0019] Palmprint feature extraction process diagram Figure 2 .

[0020] The nine areas of the left palm Figure 3 .

[0021] The nine areas of the right palm Figure 4 .

[0022] Table 1: The characteristics of the palm lines in the nine zones, possible symptoms and TCM constitutions.

Claims

1. A method for predicting physical symptoms and identifying the correspondence between TCM constitutions based on palm print features extracted by digital image processing algorithms.

2. Table 1: Palm print characteristics, possible symptoms and TCM constitutions in the nine zones.

Images