Video compression with an effective block matching algorithm and RDOT

Abstract

The storage requirement for digital video is growing day-by-day due to increase in video quality. In order to reduce the file size while maintaining the required quality, a novel compression algorithm is needed. In this paper, an Adaptive Rood Pattern Search (ARPS) is used as the Motion Estimation (ME) and Motion Compensation (MC) technique and ARPS is integrated with the proposed Rate Distortion Optimized Transform (RDOT) to reduce the computational complexity without losing the quality of the video. In this ME technique, Mean Absolute Difference (MAD) is considered as the matching criteria. ARPS technique is compared with the Diamond Search (DS) and Three Step Search (TSS) algorithms and it outperforms with respect to execution time and PSNR. Finally the combination of ARPS with RDOT achieves a high compression ratio.