A stochastic multirate signal processing approach to high-resolution signal reconstruction

This paper addresses the problem of reconstructing a signal at some high sampling rate from a set of signals sampled at a lower rate and subject to additive noise and distortion. A set of periodically time-varying filters are employed in reconstructing the underlying signal. Results are presented for a one-dimensional case involving simulated data, as well as for a two-dimensional case involving real image data where the image is processed by rows. In both cases, considerable improvement is evident after the processing.