Spatio-temporally regularized reconstruction of gated SPECT myocardial image sequences

This paper addresses the problem of improving the signal-to-noise ratio (SNR) of cardiac gated SPECT image sequences. The problem is formulated as the spatio-temporal regularization of the image sequence reconstruction using the Tikhonov framework and taking into account the image deformation. A conjugate gradient algorithm is used to compute the sequence by minimizing a quadratic objective function. The myocardial motion is estimated in a preprocessing step. A confidence coefficient is introduced for controlling locally the temporal regularization according to the accuracy of the estimated motion. Experimental results show that this approach substantially increases the signal-to-noise ratio of the reconstruction