Program Results

Introduction to the event

Current research results show that we are able to use artificial intelligence-based methods to undergo ultra-low-dose or ultra-short-time PET imaging, reducing the amount of radiotracer needed to a few percent of the original, achieving increased scanner turnover, dramatically reduced radiation dose, and increased image quality, a “win-win-win” situation (1 journal article under preparation, 1 under review, 3 published, several international/domestic conference abstracts published). In addition, as shown in the figure (published in Lin et al., IEEE T Radiol Plasma Med Sci 2025), we have successfully utilized an image fusion method to incorporate the high-resolution information from MRI into PET imaging, increasing its apparent resolution as well as proposing a novel partial volume effect correction method. Using a purely artificial intelligence deep learning-based method is also a novel approach to partial volume effect correction (1 journal paper under preparation, 1 published, several international conference abstracts published).