A Robust Least Squares Framework for PET Image Reconstruction

In this paper,we propose a robust least squares solution for real position emission tomography（PET） system to minimizing the negative effect from the system matrix uncertainty on the final image. Compared with other algorithms at present,the robust least squares method proposed takes the noise fully into consideration.The performance of the algorithm is evaluated using the simulated phantom data,and the test result shows that a significant improvement in image quality over the conventional methods of least squares reconstruction is achieved.