Computer Network Redundancy Reduction Using Video Compression

Abstract

Due to the strong correlation between symmetric frames, video signals have a high degree of temporal redundancy. Motion estimation techniques are computationally expensive and time-consuming processes used in symmetric video compression to reduce temporal redundancy. The block-matching technique is, on the other hand, the most popular and efficient of the different motion estimation and compensation techniques. Motion compensation based on the block-matching technique generally uses the minimization of either the mean square error (MSE) or mean absolute difference (MAD) in order to find the appropriate motion vector. This paper proposes to remove the highly temporally redundant information contained in each block of the video signal using the removing temporal redundancy (RTR) technique in order to improve the data rate and efficiency of the video signal. A comparison between the PSNR values of this technique and those of the JPEG video compression standard is made. As a result of its moderate memory and computation requirements, the algorithm was found to be suitable for mobile networks and embedded devices. Based on a detailed set of testing scenarios and the obtained results, it is evident that the RTR compression technique allowed a compression ratio of 22.71 and 95% loss in bit rate reduction while maintaining sufficient intact signal quality with minimized information loss.