A robust video denoising system using optimised surfacelet transform

The conventional video denoising algorithms utilizes either a strenuous motion estimation step or by the frame by frame wavelet transformation without exploiting the correlation between neighboring frames. However, these schemes of video denoising results in videos with jittery edges and curves. The limitations of motion estimation based schemes are that they suffer due to aperture problems in optical flow and lighting variations. Surfacelet transform is a powerful tool for the representation of multidimensional data. Video signals can be treated as a different type of 3D signal and therefore can be processed using surfacelet transform which preserves the edge information and visual quality. By incorporating the advantages of surfacelet transformation with an advanced optimization technique we propose a novel video denoising technique which was found to be delivering overwhelming results in terms of peak signal to noise ratio(PSNR) and structural similarity(SSIM) index.