An ai traditional chinese medicine auxiliary diagnosis and treatment system and method based on a multi-modal knowledge graph

Abstract

This invention provides an AI-assisted TCM diagnosis and treatment system and method based on a multimodal knowledge graph, belonging to the field of smart healthcare technology. It constructs a multimodal knowledge graph through a logical network between symptom entities, syndrome entities, and prescription entities, as well as a tongue visual feature library and a pulse waveform feature library. Features are extracted from tongue images and pulse signals to obtain tongue and pulse feature vectors, and semantic matching and association are performed to obtain multimodal perceptual evidence. The patient's chief complaint features are linked and traversed using the multimodal knowledge graph to obtain preliminary diagnostic hypotheses. Based on the multimodal perceptual evidence, dynamic confidence fusion and syndrome correction are performed on the preliminary diagnostic hypotheses to obtain a confidence-based diagnostic prompt. Finally, a structured TCM diagnosis and treatment plan is generated through a large language model. This invention can achieve dynamic confidence fusion of multimodal perceptual evidence such as TCM tongue diagnosis and pulse diagnosis with diagnostic knowledge logic, thereby improving the accuracy of AI-assisted TCM diagnosis.

Images