Health trend analysis method based on tongue appearance and space-time relationship and related equipment

CN121439236BActive Publication Date: 2026-07-03SHENZHEN YUANDAOMIAO MEDICAL TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHENZHEN YUANDAOMIAO MEDICAL TECH CO LTD
Filing Date
2025-12-29
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing health monitoring technologies rely on single physiological indicators or static examination data, which makes it difficult to comprehensively reflect the overall physical condition and dynamic evolution characteristics of users. Furthermore, they lack continuous analysis of physical characteristics and health trends, resulting in low accuracy and reliability of health trend analysis.

Method used

By combining traditional Chinese medicine theory with modern data technology, this study uses convolutional neural networks and deep learning methods to extract features of the tongue, complexion, and eyes, generating quantifiable physiological state feature vectors. It then uses time-series models to assess future health trends and establishes an individualized health trend analysis system.

Benefits of technology

It enables individualized and dynamic analysis of users' health status, improves the accuracy and reliability of health trend analysis, overcomes the information loss problem when a single model processes heterogeneous features, and provides a comprehensive and accurate assessment of physical health trends.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121439236B_ABST
    Figure CN121439236B_ABST
Patent Text Reader

Abstract

This application relates to the field of physiological image processing technology, providing a method and related equipment for health trend analysis based on the spatiotemporal relationship of tongue image. A baseline vector of physiological state is generated based on birth time and a Five Elements and Six Qi rule library generated according to Traditional Chinese Medicine theory. Tongue images are acquired, and tongue features are extracted. Linear processing based on a mapping function generates feature vectors corresponding to the physiological state. Facial features are extracted from facial images and linearly processed to generate feature vectors corresponding to the physiological state. Eye features are extracted from eye images to generate feature vectors corresponding to the physiological state. Based on the baseline vector, the above feature vectors are weighted to generate a health state index vector, which is input into a time-series model to obtain a health state trend assessment value. By integrating innate spatiotemporal information with acquired biomarkers, dynamic and individualized health trend analysis is achieved, effectively improving the accuracy and reliability of health trend analysis.
Need to check novelty before this filing date? Find Prior Art