A Compton scattering suppression based image reconstruction method for Digital Breast Tomosynthesis

In this paper, we present a novel statistical reconstruction method based on Compton noise suppression for Digital Breast Tomosynthesis. The physical modeling for X-ray's interaction with tissues was investigated. A Compton noise based statistical model with a conjugate prior was introduced. Results from our proposed method, Backprojection (BP) method, Simultaneous Algebraic Reconstruction Technique (SART) and Maximum Likelihood with Convex Algorithm (MLCA) were compared based on image quality analysis. Compared with other algorithms, our proposed method shows comparable performance. It also produces even better results in the Contrast to Noise Ratio (CNR) evaluation.