作者: K. V. Arya , Purnima Prasad
DOI: 10.1109/NCVPRIPG.2013.6776184
关键词: Video compression picture types 、 Block-matching algorithm 、 Multiview Video Coding 、 Computer science 、 Video tracking 、 Reference frame 、 Motion interpolation 、 Inter frame 、 Quarter-pixel motion 、 Rate–distortion optimization 、 Residual frame 、 Video denoising 、 Motion compensation 、 Frame rate 、 Data compression 、 Motion estimation 、 Video quality 、 Artificial intelligence 、 Computer vision
摘要: The popular techniques to eliminate temporal redundancy in video sequences are Motion Estimation and Compensation. These have also been used H.264, MPEG-2 MPEG-4 coding standards. Conventional fast Block Matching Algorithms (BMA) perform exhaustive search between the current reference frame. Although BMA technique gives exact result but it is computationally very expensive. Another drawback of this method that easily gets trapped into local minima which eventually lead degradation quality. proposed Technique exploits fact human eyes incapable detecting different frames when they run at particular frame rate. experimental results on various demonstrate has outperformed all existing conventional motion estimation techniques.