Video denoising using vector estimation of wavelet coefficients

摘要: Wavelet-based image denoising can be extended to a video by applying it each frame independently. The performance improved exploiting inter-frame correlations, for example, using appropriate temporal filtering. However, fixed filters might not perform sufficiently well due their inability cope with the variability of correlations across video. While many adaptive filtering approaches in spatial domain have been proposed, they do straightforwardly extend wavelet-based denoising. We propose vector extension popular hidden Markov tree modeling that flexibly exploits color and dependency wavelet coefficients. Experimental results confirm estimator coefficients yields superior existing solutions, both CPSNR visual quality sense.