List-mode image reconstruction for real-time PET imaging

摘要

The problem of tomographic image reconstruction from measured projection data can be reduced to solving a set of linear equations represented in matrix form. In this paper, we show how image reconstruction in positron emission tomography (PET) can be achieved potentially in real-time using a regularized pseudo-inverse of the system matrix. An update of the current image estimate accounting for the next registered event can be obtained using a single column of the regularized pseudo-inverse matrix, thus allowing for instant visualization of the radioactivity distribution while the subject is still being scanned. The computed image estimate is the minimum-norm least-squares solution of the regularized inverse problem. The speed and performance of the algorithm were characterized and compared to those of a well-known statistical iterative algorithm. The proposed algorithm was validated with experimental list-mode PET data.