Spatio-temporal separable data-dependent weighted average filtering for restoration of the image sequences
Kouji Miyata, Akira Taguchi · IEEE International Conference on Acoustics Speech and Signal Processing · 2002
The restoration of the image sequences corrupted by the additive noise is important for getting the high quality image and the high compression ratio of the image sequences. In this paper, we propose a novel restoration method for the image sequences corrupted by the Gaussian noise. Conventional methods use the spatio-temporal filtering after the motion compensation. However, the accuracy of the motion compensation on degraded image is not satisfied. Therefore, in the proposed method, we divide spatio-temporal filter into spatial filter and temporal filter in order to achieve the high noise reduction. The spatial filter is regarded as the pre-filter for the motion compensation. We also study the suitable filtering scheme after the motion compensation.