CN-121983248-A - AI-based traditional Chinese medicine nursing recommendation system

CN121983248ACN 121983248 ACN121983248 ACN 121983248ACN-121983248-A

Abstract

The invention relates to the technical field of intelligent traditional Chinese medicine nursing, in particular to an AI-based traditional Chinese medicine nursing recommendation system which comprises a tongue image feature extraction module, a tongue partition abnormality identification module, a meridian mapping and positioning module, a medicine property channel matching module and a nursing intervention recommendation module. According to the invention, depth feature extraction and partition quantification are carried out on the tongue image, visual information such as outlines, textures, colors and the like is converted into accurate numerical parameters, and objective judgment references are set according to statistical distribution of large-scale health samples, so that uncertainty and difference brought by traditional dependence on subjective experience evaluation are overcome, standardized identification of tongue image abnormality is realized, a calculation mode of spatial position contribution degree is introduced, tongue partition with most obvious pathological significance can be automatically focused, and further, local observation is closely related with an overall meridian system by constructing accurate mapping from tongue abnormal partition to body meridian path, and the problem of generalized fuzzy recommendation in the past is solved.

Inventors

CHENG SHI
ZHANG SHUZHEN
HOU XIAOXUE
ZHANG LIN

Assignees

南通大学

Dates

Publication Date: 20260505
Application Date: 20251208

Claims (10)

1. An AI-based traditional Chinese medicine care recommendation system is characterized in that the system comprises: the tongue image feature extraction module is used for acquiring tongue image acquisition data, extracting image features of contour edges, texture structures and color distribution of tongue tips, tongue edges and tongue root partitions, and obtaining a tongue partition feature map; The tongue body partition abnormality identification module is used for identifying abnormal types of broken outline edges, disordered texture structures and color distribution deviation of each tongue body partition in the tongue body partition characteristic diagram and generating tongue image abnormal partition annotation data; the meridian mapping and positioning module extracts abnormal part names in the tongue body partition of the tongue image abnormal partition labeling data, matches the abnormal part names with corresponding body meridian paths according to the meridian attribution relation of the traditional Chinese medicine, and outputs meridian influence path data; the medicine property channel allocation matching module refers to the meridian influence path data, screens the information of the traditional Chinese medicine materials belonging to the corresponding meridian and having the action range covering the influence path, and obtains medicine property associated data; and the nursing intervention recommending module is used for comparing the drug property related data with a preset nursing intervention knowledge base, screening nursing contents matched with the drug property characteristics and outputting traditional Chinese medicine nursing recommending results.
2. The AI-based traditional Chinese medicine nursing recommendation system as set forth in claim 1, wherein the tongue partition feature map specifically includes tongue partition identification information, image feature vector expression and feature space coordinate sets, the tongue abnormal partition labeling data includes abnormal partition positions, abnormal type labels and feature classification labels, the meridian influencing path data specifically includes meridian path names, tongue part correspondence and home organ numbers, the medicine property associated data includes medicine names, meridian number and action level information, and the traditional Chinese medicine nursing recommendation result specifically refers to nursing action names and intervention part positioning.
3. The AI-based traditional Chinese medicine care recommendation system of claim 1, wherein the tongue image feature extraction module comprises: The image feature initial extraction sub-module is used for acquiring tongue image acquisition data, inputting the tongue image acquisition data into a residual convolution neural network with a cross-layer connection structure, extracting features of parallel contour edges, texture structures and color distribution of images, and carrying out numerical dimension calculation on the extracted multi-channel features to obtain multi-dimensional feature vectors; the partition characteristic marking submodule divides the multidimensional characteristic vector into corresponding partitions according to the preset coordinate masks of tongue tip, tongue edge, tongue middle and tongue root partitions, calculates the curvature value of the contour edge in each partition, the complexity of texture structures and the average value of color distribution, and generates a tongue domain characteristic parameter set; And a feature map fusion construction submodule for carrying out multichannel result integration on the tongue domain feature parameter set and establishing a tongue body partition feature map according to the mapping arrangement of the space position relation of the partition.
4. The AI-based traditional Chinese medicine care recommendation system of claim 1, wherein the tongue partition abnormality identification module comprises: The space feature calculation sub-module calls the tongue body partition feature map, calculates the space position feature of each partition of the tongue tip, the tongue edge, the tongue and the tongue root through a position attention mechanism function, and quantifies the contribution degree of the contour edge, the texture structure and the color distribution feature points in the partition to be used as partition space attention weight; The partition abnormality judging sub-module is used for comparing the contour edge fracture, the texture structure disorder and the color distribution deviation characteristic of each partition in the tongue partition characteristic diagram with a preset abnormality judging reference value item by item based on the partition space attention weight, and executing partition abnormality type judgment to obtain a partition abnormality class set; The abnormal partition marking sub-module marks the identified partition abnormal class set and the corresponding profile edge fracture, texture structure disorder and color distribution deviation characteristics, and associates and sorts the abnormal class set with class information of tongue tips, tongue edges and tongue root partitions in the tongue, so as to generate tongue image abnormal partition marking data.
5. The AI-based traditional Chinese medicine care recommendation system of claim 1, wherein the meridian map locating module comprises: The abnormal part extraction sub-module is used for identifying and extracting the abnormal part names marked in the tongue image abnormal partition marking data, and performing duplicate removal operation on the repeatedly-occurring names to obtain an abnormal part name list; The meridian relation matching sub-module calls the meridian belonging relation of the traditional Chinese medicine according to the abnormal part name list, maps the heart meridian according to the tongue tip, maps the liver and gall meridian according to the tongue edge, maps the spleen and stomach meridian in the tongue, maps the kidney meridian and bladder meridian according to the tongue root, performs meridian mapping operation on each part name in the list, and establishes a meridian mapping relation set; and the influence path construction submodule is used for associating each abnormal part name in the meridian mapping relation set with the matched body meridian path, constructing a corresponding relation between the tongue image and the meridian, and obtaining meridian influence path data.
6. The AI-based traditional Chinese medicine care recommendation system as claimed in claim 5, wherein the meridian attribution relationship of the traditional Chinese medicine is a preset structured lookup table, the structured lookup table comprises two data fields of tongue body partition and meridian path, and key value pairs of tongue tip and heart meridian, tongue edge and liver and gall meridian, tongue middle and spleen stomach meridian, tongue root and kidney meridian and bladder meridian are stored in advance.
7. The AI-based traditional Chinese medicine care recommendation system of claim 1, wherein the drug property menstruation-regulating matching module comprises: The meridian path analysis submodule is used for executing structural analysis on path information in the meridian influence path data, extracting meridian names and corresponding organ types, and summarizing the identified meridian names and organ types to obtain a meridian organ list; the medicinal material information screening submodule compares the meridian names in the meridian organ list with meridian returning attributes of all medicinal materials in a medicinal herb meridian returning information base, screens the medicinal material items of the corresponding meridians and collaterals, and establishes a candidate medicinal material information set; and the medicine property characteristic matching sub-module is used for judging whether the meridian direction and the action level of each medicine accord with the requirements in the influence path based on the candidate medicine information set and the meridian influence path data, extracting medicine information of which the action range can cover the influence path, determining medicine property characteristics related to the current tongue picture and generating medicine property related data.
8. The AI-based traditional Chinese medicine care recommendation system of claim 7, wherein the meridian tropism attribute is a text set composed of a plurality of standardized meridian names, and the text set is compared with the meridian names in the meridian organ list to judge the matching of the meridian tropism.
9. The AI-based traditional Chinese medicine care recommendation system of claim 1, wherein the care intervention recommendation module comprises: the drug property characteristic analysis submodule is used for carrying out structural analysis on the drug property characteristics and the action range in the drug property associated data, identifying the needed conditioning direction and the body intervention part and generating a conditioning requirement list; The nursing content screening sub-module is used for comparing the applicable conditioning direction and the intervention area of each item of nursing content in a preset nursing intervention knowledge base with the requirements in the conditioning requirement list, screening the nursing content matched with the drug property characteristics and establishing a matched nursing item set; and the recommended result generation sub-module is used for sequencing nursing action contents in the items according to the association relation between the medicine property conditioning range and the nursing action intervention area based on the matched nursing item set and outputting traditional Chinese medicine nursing recommended results.
10. The AI based Chinese medicinal care recommendation system of claim 9 wherein the applicable conditioning direction is a text set of a plurality of conditioning direction keywords, the text set is compared to the conditioning requirement list as an input for a match determination, the intervention region is a text set of a plurality of body part names, and the text set is matched to the body intervention parts in the conditioning requirement list during the comparison operation.

Description

AI-based traditional Chinese medicine nursing recommendation system Technical Field The invention relates to the technical field of intelligent traditional Chinese medicine nursing, in particular to an AI-based traditional Chinese medicine nursing recommendation system. Background The technical field of intelligent traditional Chinese medicine nursing relates to the combination of traditional Chinese medicine nursing theory and modern informatization means, and core matters comprise the formulation and adjustment of traditional Chinese medicine nursing schemes, compatibility selection of traditional Chinese medicine formulas and nursing measures, dialectical basis application in the nursing process and acquisition and utilization of individual health data. The traditional Chinese medicine nursing recommendation system is a tool for selecting a proper traditional Chinese medicine nursing scheme based on basic health information of a patient and the key points of traditional Chinese medicine differentiation, and generally aims at technical matters such as symptom expression, tongue pulse characteristics, physique type and the like of the patient. The prior art processes the symptoms of patients and the symptoms of tongue pulse characteristics and other differentiation points, which are often dependent on qualitative descriptions of manual observation and empirical judgment, the mode is easily influenced by subjective factors of operators to cause lack of consistency and repeatability of recommended results, meanwhile, the analysis of tongue images is often macroscopic judgment of the integrity, the subtle anomalies of specific areas of the tongue body, such as yellow tongue coating, cannot be precisely quantified and positioned, the pathological meanings represented by the tip or the root of the tongue are quite different, the prior art is difficult to distinguish the spatial differences, so that the precise correspondence between the abnormal parts and specific channels and collaterals cannot be established, the generated nursing scheme is not strong in pertinence, in addition, the selection process from the differentiation points to the nursing scheme is often a simple and preset macroscopic match, the intervention area from the pathological characteristics to the drug properties is lack of channel, the layering of the nursing scheme is performed, the refined matching logic is caused, the correlation degree between the recommended nursing content and the specific affected parts and the nursing requirements of the body is insufficient, for example, the nursing content with the effect of promoting dampness is not recommended according to the wet heat, the nursing content is really indicated, the important nursing scheme is not to act on the relevant specific channels and collaterals, and collaterals are difficult to realize the precise and the specific intervention on the body is difficult to realize. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides an AI-based traditional Chinese medicine nursing recommendation system. In order to achieve the purpose, the invention adopts the following technical scheme that the traditional Chinese medicine nursing recommendation system based on AI comprises: the tongue image feature extraction module is used for acquiring tongue image acquisition data, extracting image features of contour edges, texture structures and color distribution of tongue tips, tongue edges and tongue root partitions, and obtaining a tongue partition feature map; The tongue body partition abnormality identification module is used for identifying abnormal types of broken outline edges, disordered texture structures and color distribution deviation of each tongue body partition in the tongue body partition characteristic diagram and generating tongue image abnormal partition annotation data; the meridian mapping and positioning module extracts abnormal part names in the tongue body partition of the tongue image abnormal partition labeling data, matches the abnormal part names with corresponding body meridian paths according to the meridian attribution relation of the traditional Chinese medicine, and outputs meridian influence path data; the medicine property channel allocation matching module refers to the meridian influence path data, screens the information of the traditional Chinese medicine materials belonging to the corresponding meridian and having the action range covering the influence path, and obtains medicine property associated data; and the nursing intervention recommending module is used for comparing the drug property related data with a preset nursing intervention knowledge base, screening nursing contents matched with the drug property characteristics and outputting traditional Chinese medicine nursing recommending results. The invention improves that the tongue body partition characteristic map specifically comprises tongue body partition