PET image improvement using the Patch Confidence K-Nearest Neighbors Filter

摘要

In Positron Emission Tomography (PET), the resulted images are highly deteriorated by noise. In this study, we propose a new denoising framework using the Patch Confidence K-Nearest Neighbors Filter (PCKNN) to reduce noise in the sinogram before forwarding it to the reconstruction procedure. This method has been evaluated on a simulated PET image of a phantom, and the performance has been compared with several conventional methods in the literature. The results have shown that the PET image quality can be substantially improved in term of increased signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR).