An efficient wavelet based video compress system based on the novel motion estimation algorithm

摘要

Image and video coding performs a significant task in several multimedia applications. In recent years, digital video communication can be found in numerous applications like broadcast services over satellite and terrestrial channels, digital video storage, wired and wireless conversational services and many more. Without processing, it is not possible to store the complete digital video. Video compression denotes the process of decreasing the amount of data required for representing the digital video images and it is the combination of spatial image compression and temporal motion compensation. So far, a lot of compression algorithms had been developed for reducing the data quantity and to render more satisfactory video quality. The main objective for carrying out this research is to construct an efficient video compression system. There are three stages involved in the proposed video compression system. In the first step of video compression, wavelet decomposition is applied to the I-frame and the resulting coefficients are quantized by employing the Listless SPECK (LSK) algorithm. The various kinds of wavelets used in the proposed LSK compression algorithm are the Daubechies, the Coiflet, the Biorthogonal and the Reverse Biorthogonal. The compression ratio gets effectively modified with different types of wavelets used. A novel block motion estimation algorithm is utilized for estimating the motion between the various video blocks. The comparison on the performance of LSK algorithm that uses different kinds of wavelets were made and the performance measures used to analyze the algorithm are the Compression Ratio (CR) and Peak Signal to Noise Ratio (PSNR). An effective estimate of the motion in the frames is obtainable with these performance measures. A highly efficient and effective video compression is achieved with the proposed scheme. Since LSK has been applied to different wavelet decompositions, the algorithm is more advantageous in terms of speed and simple hardware implementation.