Adaptive-centre candidate decimation distance-dependent thresholding search for motion estimation
Sorwar, G 2005, 'Adaptive-centre candidate decimation distance-dependent thresholding search for motion estimation', Proceedings of 3rd International Conference on Information Technology and Applications, ICITA 2005, Sydney, NSW, 4-7 July, IEEE, pp. 674-679.
Published version available from:
Fast Motion estimation algorithms for compression of moving pictures are widely required for real time video encoding instead of the exhaustive block motion estimation algorithm, known as the full search (FS) algorithm which provides optimal error performance but requires enormous computation for calculating motion vectors. The author has previously proposed a motion estimation algorithm, namely adaptive centre distance-dependent thresholding search (ACDTS) by introducing the concept of a distance-dependent thresholding search (DTS) with adaptively predict the starting search centre for fast and flexible motion estimation for video coding applications. Though, the performance of ACDTS algorithm was well for low motion video sequences, it was not so satisfactory for high motion video sequences. In this paper, the ACDTS algorithm has been modified into adaptive-centre candidate decimation distance-dependent thresholding search (ACDDTS) algorithm by incorporating a nonlinear centre biased search point pattern with ACDTS algorithm for trading off quality and processing speed. Experimental results show that the proposed ACDDTS algorithm outperforms the well known fast three step search (TSS), new three step search (NTSS), and adaptive centre new three step search (ACNTSS) algorithms by reducing the computational time with all motion types video sequences.