Mutual Noise Estimation Algorithm for Video Denoising
Mohammad Izadi, Neil Birkbeck, Balu Adsumilli · 2019
This paper presents a novel algorithm to estimate spatio-temporal noise variance in videos. The algorithm uses mutual information from the spatial and temporal noise statistics to detect homogeneous blocks in a video frame and estimate the noise. The experimental results show the accuracy and robustness of the algorithm over videos with low to high spatial and temporal complexities, and varying levels of noise power and noise correlations.