Implementing of maximum likelihood in tomographical coded aperture

We performed investigations in coded aperture to improve the spatial and depth resolution. Coded aperture imaging has demonstrated a good efficacy, specially concerning its sensitivity. It consists in photon emission encoding through a specific collimator. Such a collimator includes several pinholes distributed on an opaque lead plate. The pattern of hole distribution used is based on a Singer array. Decoding is by reconstruction of nine tomographic slices from one single acquisition. The maximum likelihood method (MLE) was used. In this approach, acquired measurements are considered as random variables. The maximization of the likelihood function is obtained by the expectation-maximization (EM) algorithm. Our study gave the following preliminary results. The maximum likelihood method improved the spatial resolution of the slices. The noise introduced by the reconstruction process is also clearly reduced. The depth resolution improved but was still not sufficient for clinical imaging. Our results underline the imaging problems resulting from a restricted view angle acquisition.