A NOVEL VIDEO CODING FRAMEWORK BASED ON COMPRESSED SENSING AND DISCRETE WAVELET CODING APPROACH FOR EDGE PRESERVING

Abstract Nowadays people have to handle a large amount of data at every time which is even highly not easy with the increasing development of technology and the opening into the new trend of digital age. So there is a necessity to store and retrieve the digital information in a scalable and better way which is highly useful for practical use. With this motivation wavelets used to provide a mathematical way of encoding the data recently in such a way that according the level of details it is layered. This layering helps approximations at various intermediate stages and these approximations can be accumulated with a lot less space when compared to the original information.In this research work, an efficient video compression framework is proposed based on compressed sensing in the wavelet domain. The new framework contains four stages in which the source images are signified with their sparse coefficients by using the Discrete Wavelet Transform (DWT) initially. In the next stage, usingthe random Gaussian matrix the measurements are attained from their sparse coefficients. Thirdly novel Fast Continuous Linearised Augmented Lagrangian Method (FCLALM) is utilized to reconstruct the sparse coefficients that will be transformed by the Inverse DWT (IDWT) to the fused image. Finally Local Linear Stein's Unbiased Risk Surface Estimator (LLSURE) filter is proposed for edge-preserving surface estimator. The proposed estimator is used to leave some noise in the vicinity of edges and improve those edges during denoising process. The experimental results demonstrate that the LLSURE with FCLALM has a provides higherPeak-Signal- To-Noise Ratio(PSNR), Lesser Mean Square Error(MSE), and a faster convergence rate when compared with other some other existing edge preserving methods such as LLSURE, SURE and Optimal Compression Plane (OCP).