Spine health monitoring method and system combining remote sensing three-dimensional modeling and AI pre-diagnosis

Abstract

The invention provides a spinal health monitoring method and system combining remote sensing three-dimensional modeling and AI pre-diagnosis. The method comprises the following steps: (1) constructing a rapid three-dimensional reconstruction module based on flow diffusion: realizing high-fidelity three-dimensional modeling by adopting a lightweight shape generation and intelligent texture mapping technology; (2) intelligent grading evaluation of scoliosis driven by a large model: based on a large-scale spine feature data set, constructing a deep learning network for automatic measurement and typing prediction of spine parameters, and deeply fusing the deep learning network with a semantic understanding module of an AI intelligent question-answering system; and (3) doctor-worker collaborative diagnosis and feedback optimization: doctors can view the user model and the AI evaluation report and make new diagnosis, meanwhile, the system utilizes natural language processing to extract doctor diagnosis keywords, and the model is continuously optimized based on parameter transfer learning. According to the invention, the whole-process Web deployment of spine detection is realized, a medical vertical domain large model for spine detection is developed, and the extensible technical framework has huge application potential in the field of intelligent medical treatment.

Images