Construction of bisection model of SPECT bone scan image based on VGGNet

Abstract

Nuclear medical SPECT imaging is an advanced medical imaging equipment, which plays an important role in the discovery, diagnosis and treatment of bone metastases in the process of medical diagnosis. Computerized SPECT imaging can accurately diagnose whether patients have bone metastasis, which can help doctors quickly identify whether there is disease. In this paper, the whole body bone scanning imaging data is effectively expanded by mirroring, translating and rotating the existing data, and then an image classifier is constructed based on VGGNet model. The experimental evaluation and analysis of a group of real tomographic data by image classifier shows that VGGNet7 model can effectively distinguish disease from normal, and the accuracy Acc, PRecision pre, recall rec and F-1 scores of experimental evaluation are 0.99, 0.99, 0.99 and 0.99, respectively.