Name
Abstract | ||
|---|---|---|
The H.264 [1] video coding standard achieves highest coding gain. In addition to common tools in previous standards such as H.263 and MPEG4, it supports multiple frames, multiple block sizes, and 1/4 pixel motion estimation, deblocking filter, intra-prediction, and so on. However, full search of for all block modes and reference frames requires heavy computational complexity. The algorithm Predictive Motion Vector Field Adaptive Search Technique (PMVFAST) [3] was previously accepted into MPEG standard to achieve hundreds of time of speed up. FMBME [4], and some other algorithms [5]-[8] have been developed to do fast multiple block size motion estimation. For multiple reference frame motion estimation, FMFME [9], and some other algorithms [10] - [11] have also been developed. In this paper a new algorithm is presented named as Fast Mode Decision and Motion Estimation (FMDME). This algorithm achieves high speed up factor by involving Fast SKIP mode checking, Fast multiple block size motion estimation, Fast multiple reference frame selection, Fast Motion Estimation, and Fast Intra-block skipping for P frame in H.264. |
Year | DOI | Venue |
|---|---|---|
2006 | 10.1109/ISCAS.2006.1692625 | 2006 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-11, PROCEEDINGS |
Keywords | Field | DocType |
computational complexity,coding gain,reference frame,gold,motion estimation,prediction algorithms | Reference frame,Block size,Computer vision,Block-matching algorithm,Quarter-pixel motion,Motion detection,Computer science,Artificial intelligence,Motion analysis,Motion estimation,Deblocking filter | Conference |
ISSN | Citations | PageRank |
0271-4302 | 3 | 0.44 |
References | Authors | |
3 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Hoi-ming Wong | 1 | 68 | 8.76 |
| Oscar C. Au | 2 | 1592 | 176.54 |
| Andy Chang | 3 | 3 | 0.44 |
| Shu-kei Yip | 4 | 60 | 7.43 |
| Chi-wang Ho | 5 | 57 | 7.35 |