An unsymmetrical-cross multi-resolution motion search algorithm for MPEG4-AVC/H.264 coding

The new H.264 (MPEG-4 AVC) video-coding standard has a significant performance benefit compared to former standards. Unfortunately, those advanced coding features, including variable block-size motion compensation and multiple reference frames, incur a considerable increase in encoder complexity, especially when using the straightforward full search (FS) algorithm, mainly with regards to motion estimation (ME) and mode decision. We propose a new ME algorithm for fast H.264 coding, utilizing the technique of the generalized motion vector (MV) predictors, named hybrid unsymmetrical-cross multi-resolution grid search (UCMRGS), which takes advantage of the correlation among variable blocks and can considerably reduce the computational cost of ME at the encoder, while at the same time give similar, and, in some cases, better, visual quality compared with the brute force full search algorithm. The proposed algorithms mainly rely upon very robust and reliable predictive techniques with parameters adapted to the local characteristics combined with the UCMRGS pattern