Recursive Type of Video-Data Dependent Weighted Average Filters for Video Signals Corrupted by Gaussian Noise

The restoration of video signals corrupted by the additive noise is important for getting the high quality image and the high compression ratio of video signals. In this paper, we propose a novel restoration method for video signals corrupted by the Gaussian noise. In this paper we propose a recursive type of spatio-temporal filter. Conventional spatio-temporal filters are used only input sequences (i.e., 3D-FIR filters). 3D-FIR filters cannot be obtained good filtering performance in still regions of an image sequence. On the other hand, the recursive type filters have strong noise attenuation property at the still region. However, the motion is degraded by the recursive type filtering. To overcome this drawback, we add motion information that controls the weight of the previous output sequence. Thus, the proposed data dependent filter can reduce the additive noise efficiently while preserving the signal edge/detail and motion