Spatial-Temporal Filtering for Video Denoising using Weighted Highpass Filtering Coefficients

Abstract: This paper proposed a novel spatial-temporal filtering method for video denoising in wavelet domain. In proposed video denoising algorithm, spatial adaptive noise filtering in wavelet domain is combined with temporal filtering in time domain. Spatial filtering of individual frames is done by taking discrete wavelet transform (DWT) and modified wavelet coefficients by Weighted Highpass Filtering Coefficients (WHFC). Further apply adaptive wiener filter to the reconstructed frames. But, the denoising artifacts and the remaining noise differed from frame to frame and produce unpleasant visual effects. So temporal filtering is essential. In our method modified spatial filtering is associated with temporal filtering, which is based on block based motion detector and on selective recursive time averaging of frames. The performance of proposed algorithm is evaluated in terms of Peak Signal to Noise Ratio (PSNR) and Structural Similarity Index Metric (SSIM). Experiment results show that the proposed algorithm gives higher PSNR as well as better visual quality.

Keywords: DWT, Adaptive wiener filter, WHFC, PSNR and SSIM.