Edge-adaptive Kalman filtering for image restoration with ringing suppression

Abstract

The authors extend the two-dimensional (2-D) linear space-invariant (LSI) reduced update Kalman filter (RUKF) to edge-adaptive space-invariant restoration of noisy and blurred images using a decision-directed approach. The edge-adaptive RUKF was motivated by the need to suppress the ringing artifacts caused by LSI processing. The authors show that ringing artifacts can be suppressed to a great extent by using multiple image models that provide a better match to local edge orientation. The maximum a posteriori probability decision procedure developed for model selection at each pixel can be used with other 2-D Kalman filtering algorithms as well.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>