Wiener and Kalman Filters for Denoising Video Signals

In this report, we are investigating the performance of Wiener & Kalman filters when applied to the problem of denoising video signals. When applying Wiener filer we define the neighborhood around the pixel either frame by frame, spatial-temporal neighborhood or use the matching block from previous frame. We also explain how to model video signal so that Kalman filter assumptions hold true. We try varying the size of the neighborhood and then compare our results to other conventional denoising filters like median filter and nonlinear soft coring technique. We show that Wiener filter results are comparable to median filter and that Kalman filter is performing the best with some blocking artifacts