Detail preserving noise filtering for compressed image

While high compression ratio has been achieved using recently developed image coding algorithm, the noise removal technique is considered as an important subject. This still holds for very low bitrate video coding, that is, MPEG-4 has defined it as a core experiment which is mainly concerned with block based discrete cosine transform (DCT) coding such as H.263 and MPEG-1. This paper describes a novel and practical technique which attempts to accomplish both noise suppression and detail preservation at the same rime. Some of the conventional adaptive filters are designed to search a homogeneous region among the predetermined polygonal subregions, then to apply a smoothing operation within the selected subregion. It shall be, however, noted that sometimes the predetermined subregion finally selected may still be heterogeneous. This fact leads us to a novel idea; instead of examining the predetermined regions, define a lot more flexible region likely to be homogeneous. In order to achieve this, we introduce the binary index: each pixel is classified into either the lower intensity group or higher intensity group based on a local statistics. Then a smoothing operation is applied within the pixels having the same group index as the pixel to be processed. Thus our scheme can search a homogeneous region appropriately. The adaptive smoothing adopted in the proposed scheme is also designed to be consistent with an important property of human visual system, i.e., the spatial masking: noise visibility decreases at spatial details such as edges and textures. Another advantage is that it can be realized with significantly low computations. The simulation results show thar this approach can suppress the visible artifacts while retaining the fine derails such as edge and texture.