科研之友
微信
新浪微博
Facebook
分享链接
摘要
In this paper, we present a novel image reconstruction algorithm for positron emission tomography(PET). Almost all of existing reconstruction approaches assume that the measurement model for PET is linear equation with Gaussian white noise or energy-bounded noise, which only approximates the emission and detection of PET very roughly. In fact, the real situation is much more complicated than the one mentioned above and there must be something that is not be involved in the aforementioned model. Hence, in this paper, we establish a more general and vivid measurement model via involving an unknown input, and propose a reconstruction method based on the optimal filtering for the stochastic system with unknown input. The approach reconstructs the PET image effectively and its performance is evaluated with the computer-synthesized cardiac-phantom.
