Color video denoising based on interframe and intercolor correlation with LMMSE filtering
B. Indhu, C. Saraswathy · 2014
The denoising method is to remove the additive white Gaussian noise in the color video sequences based on interframe and intercolor correlation prediction is presented. Noises are added uniformly and nonuniformly then the temporal correspondence is achieved through the joint-RGB motion estimation to estimate motion between the frames. Pure interframe, pure intercolor and combined interframe-intercolor correlations are predicted and compared between them. Linear minimum mean square error filtering (LMMSE) is applied to the current noisy observation as well as the interframe and intercolor prediction to estimate every noise free color component. Finally the denoised performance are evaluated by the quality metric, color peak signal to noise ratio (cPSNR).