Minimum deconvolution and 0 - 90° subset, new algorithms for emission tomography with large hole collimator

Radionuclide imaging is right now one of the poorest medical imaging modalities (MIM). With a sensitivity of 10
^-4
and a 1-1.5 cm spatial resolution it presents the worst performance. Its usefulness must be great to survive beside the others MIM. Emission tomography with large hole collimator was proposed to improve these figures. Interesting results has already been obtained but a robust algorithm is needed to give a practical proof of the superiority of the approach. This paper unveils new ways to deconvolute and reconstruct these tomographic data, both in noise free and in low photon counts conditions. Comparison with thin parallel hole collimator is also depicts.