Recursive adaptive Wiener filtering

One important image processing task concerns the restoration of blurred images degraded by additive noise. This paper describes and compares locally adaptive Wiener filtering techniques in the spatial domain. An exponentially decaying autocorrelation function is assumed, and a noncausal filter is developed whose adaptive properties are dependent on the local signal autocorrelation. The development yields a recursive filter with pole positions based on local signal and noise variance and local signal autocorrelation. Synthetic and real images are used to demonstrate the adaptive nature. The results show that the filters developed provide effective smoothing due to a large region of support, reasonable edge preservation especially in noisy and low contrast conditions, and smoothing along edges. The mean squared error is less than that of other noncasual Wiener filters and less than that of the Lee filter.