Regularized Emission Image Reconstruction Using Imperfect Side Information

Abstract

A spatially variant penalized-likelihood method for tomographic image reconstruction based on a weighted Gibbs penalty was investigated. The penalty weights are determined from structural side information, such as the locations of anatomical boundaries in high-resolution magnetic resonance images. Such side information will be imperfect in practice, and a simple simulation demonstrated the importance of accounting for the errors in boundary locations. Methods are discussed for prescribing the penalty weights when the side information is noisy. Simulation results suggest that even imperfect side information is useful for guiding spatially variant regularization.