Adaptive weighted dα filter

In this paper we propose a new adaptive weighted dα filter. The filter is adaptive regarding noise amplitude distribution, orientation of structures and anisotropy measures. The filter coefficient are chosen according to structure orientation and anisotropy measures. α value is chosen according to the result of local noise distribution and anisotropy coefficient estimations. Some experimental results on synthetic and natural images are presented. Results are compared with those of adaptive filters such as the adaptive trimmed mean.