Impulse response analysis for several digital tomosynthesis mammography reconstruction algorithms

ABSTRACT Digital tomosynthesis mammography algorithms allow reconstructions of arbitrary planes in the breast from limitedangle series of projection images as the x-ray source moves along an arc above the breast. Though several tomosynthesis algorithms have been proposed, no complete comparison of the methods has previously been conducted. This paper presents an analysis of impulse response for four different tomosynthesis mammography reconstruction algorithms. Simulated impulses at different 3-D locations were simulated to investigate the sharpness of reconstructed in-plane structures and to see how effective each algorithm is at removing out-of-plane blur. Datasets with 41, 21 and 11 projection images of the impulse were generated with a total angular movement of +/10 degrees of the simulated x-ray point source. Four algorithms, including shift-and-add method, Niklason algorithm, filtered back projection (FBP), and matrix inversion tomosynthesis (MITS) are investigated. Compared with shift-and-add algorithm and Niklason method, MITS and FBP performed better for in-plane response and outof-plane blur removal. MITS showed better out-of-plane blur removal in general. MITS and FBP performed better when projection numbers increase.