Nonstationary noise locally adaptive filtering in DCT domain

In many image processing applications, observed images are contaminated by a nonstationary noise and no a priori information on noise statistics is available. In order to remove such a noise, a locally adaptive filter has to be applied. In this paper, we propose a locally adaptive filter based on evaluation of image local activity in a “blind” manner and on discrete cosine transform (DCT) computed in overlapping blocks. The filter local adaptation takes into account the local estimation of a noise variance in image blocks and exploits discrimination of locally homogeneous and heterogeneous image regions by adaptive threshold setting. The designed filter performance is tested for simulated data as well as for real life images. Recommendations concerning filter parameter setting are provided.