Except for single-mode traditional Chinese medicine imaging techniques such as infrared thermal imaging, the one-stop multimodal whole-body imaging health examination technology and device is still blank. We focus on infrared thermal imaging as the main modality, integrated various modalities of medical imaging intelligent sensing agents such as terahertz imaging. The upper and lower computer and virtual instrument architecture are used, and the imaging data are collected by the lower computers that each is an intelligent sensing agent. The upper computer is used for image reconstruction with intelligent algorithms. Based on the core theory of traditional Chinese medicine, intelligent fusion imaging is achieved through various modalities to achieve the ‘observation, hearing, questioning, and palpation’ four diagnostic integration. We use fractional Fourier transform to filter imaging data, Laplacian pyramid for image fusion. We have proposed an implementation method and process for combining traditional Chinese medicine imaging large language model with knowledge graph, and based on deep learning, we have studied the image and report generation algorithm that combines traditional Chinese medicine pathology and four diagnostic methods with knowledge graph fusion, as well as the traditional Chinese medicine human physiological and pathological interpretation and evaluation system. We have achieved some results, and through further research and development, we can achieve commercial applications.
Read full abstract