Fast Global Motion Estimation

Global motion estimation (GME) plays an important role in many video application systems such as video coding system MPEG-4. However, its computational complexity is very high. Fast algorithms are needed. In this paper, we propose an improvement to the GME algorithm. We achieve this by introduce two techniques. Firstly, integral projection algorithm (IPA) is used to get first translation estimation and reduce block matching algorithm (BMA) search range. Secondly, coarsely block sampling technology directly reduces the computation complexity. GME composite performance experiments prove that our GME method only degrades −0.13dB in PSNR compared with the MPEG-4 verification model (VM), while it gets a 5.93ms per frame speed and is 73.3 times faster than the VM. Global motion compensation (GMC) coding experiments also show no loss of GME accuracy and compression efficiency compared to the MPEG-4 VM GME method.

[1]  Ezzatollah Salari,et al.  Successive elimination algorithm for motion estimation , 1995, IEEE Trans. Image Process..

[2]  T Koga,et al.  MOTION COMPENSATED INTER-FRAME CODING FOR VIDEO CONFERENCING , 1981 .

[3]  Mohammed Ghazal,et al.  Robust Global Motion Estimation Oriented to Video Object Segmentation , 2008, IEEE Transactions on Image Processing.

[4]  Frédéric Dufaux,et al.  Efficient, robust, and fast global motion estimation for video coding , 2000, IEEE Trans. Image Process..

[5]  Hussein R. Al-Zoubi,et al.  Reducing the complexity of MPEG-4 global motion estimation using pixel subsampling , 2008 .

[6]  A. Kak.  Computerized tomography with X-ray, emission, and ultrasound sources , 1979, Proceedings of the IEEE.

[7]  Amir Averbuch,et al.  Fast gradient methods based global motion estimation for video compression , 2002, Proceedings. International Conference on Image Processing.

[8]  Atsushi Morimura,et al.  Automatic image stabilizing system by full-digital signal processing , 1990 .

[9]  Amir Averbuch,et al.  Fast gradient methods based on global motion estimation for video compression , 2003, IEEE Trans. Circuits Syst. Video Technol..

[10]  Xiqi Gao,et al.  A multilevel successive elimination algorithm for block matching motion estimation , 2000, IEEE Trans. Image Process..

[11]  Murat Kunt,et al.  A new two-stage global/local motion estimation based on a background/foreground segmentation , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

[12]  W. David Pan,et al.  Very Fast Global Motion Estimation using Partial Data , 2007, 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07.

[13]  Bin Qi,et al.  Robust and fast global motion estimation oriented to video object segmentation , 2005, IEEE International Conference on Image Processing 2005.

[14]  W. David Pan,et al.  Efficient Global Motion Estimation using Fixed and Random Subsampling Patterns , 2007, 2007 IEEE International Conference on Image Processing.

[15]  Hiroshi Watanabe,et al.  Two-stage motion compensation using adaptive global MC and local affine MC , 1997, IEEE Trans. Circuits Syst. Video Technol..

[16]  Peyman Milanfar.  A model of the effect of image motion in the Radon transform domain , 1999, IEEE Trans. Image Process..

[17]  Noboru Wakami,et al.  An application of fuzzy set theory for an electronic video camera image stabilizer , 1995, IEEE Trans. Fuzzy Syst..

[18]  Gagan B. Rath,et al.  Iterative least squares and compression based estimations for a four-parameter linear global motion model and global motion compensation , 1999, IEEE Trans. Circuits Syst. Video Technol..