Fast adaptive motion estimation search algorithm for H.264 encoder

• Main Author: Patwary, Md Anwarul Kaium
• Format: Thesis
• Language: English
• Published: 2012
• Online Access: http://psasir.upm.edu.my/id/eprint/30925/1/FSKTM%202012%203R.pdf; http://psasir.upm.edu.my/id/eprint/30925/

The latest H.264/AVC encoder adopted more advanced techniques such as multiple reference-frame motion estimation, 4 x 4 integers Discrete Cosine Transform (DCT), intra prediction, de-blocking filter, quarter pixel Motion Estimation (ME) with variable block size and novel entropy. Motion estimation is a technique of video compression and video processing applications; it extracts motion information from the video sequence. Multiple reference-frame motion estimation can gain better compression efficiency of video coding for H.264 than previous video standards (e.g MPEG-2,H.263, JPEG). But it leads to higher computational cost and complexity in coding. In this study we proposed an efficient early termination searching method to reduce the computational complexity and achieve better compression ratio. Adaptive search strategy is applied to reduce the search point in a search range. Furthermore this study presents an analysis of the performance of the proposed algorithm in terms of motion estimation time, total encoding time, video quality (PSNR), and bit rate. Simulation result shows that as compared to previous research, this algorithm achieves up to average 60% reduction in motion estimation time without degrading the video quality.