Medical imaging is widely used in clinical diagnosis, treatment and prognosis evaluation as a non-invasive technology. In recent years, artificial intelligence has promoted new progress in the field of medical imaging. Multimodality image fusion techniques further increase the application and clinical value of medical images through richer and more comprehensive data mining and analysis. This report will focus on the application of multimodality fusion techniques for PET, CT and others in specific problems such as image reconstruction, object detection and chronic disease screening, and discuss the possible ideas, challenges and opportunities in the in-depth combination of artificial intelligence technology and classical physical principles of medical imaging.