Spatially Adaptive Temporal Smoothing for Reconstruction of Dynamic Image Sequences

In this paper, we propose a method for spatio-temporal reconstruction of dynamic image sequences. In a method we proposed previously, temporal smoothing in a Karhunen-Loegraveve (KL) or principal components (PC) transform domain was used prior to reconstruction to reduce the effect of noise. Unlike the Bayesian priors that are usually used in image reconstruction, temporal KL smoothing is a data-driven approach that takes advantage of the fact that the desired part of the data is characterized by strong interframe correlations, whereas the noise is uncorrelated. A potential disadvantage of KL-based methods is that they typically use a pooled estimate of the signal covariance matrix, thus assuming that all pixels obey similar time functions. In this paper, we investigate the possibility of making the temporal smoothing adapt spatially to local characteristics in the projection data. This can improve the noise performance of the temporal smoothing, while lessening the possibility of signal distortion. Computer simulation results are used to evaluate the technique for dynamic imaging applications in brain and tumor imaging