Power efficient low bit-rate motion estimation and compensation techniques

The block-based model for motion estimation and compensation forms the basis of almost all video standards, for example MPEG-1, MPEG-2 and H.26x. Recently, the mesh-based model has been introduced in current video and graphics standards like MPEG-4 and VRML. It offers many advantages over the block-based model. Therefore, a two-stage global motion estimation and compensation technique is proposed in this dissertation.; The first stage utilizes a new low bit-rate motion estimation and compensation techniques based on a new generalized quadtree region-wise structure. This technique presents a new segmentation procedure of video frames that enhances the compression ratio and decreases the blocking artifact factor by capturing motion correlation in a spatio-temporal dimension. The proposed encoding schemes for the generalized quadtree along with the proposed temporal predictive coding result in achieving very low bit-rate with better video quality.; An active mesh-based motion estimation and compensation technique is also developed and utilized in the second stage. It utilizes the motion information provided by the segmentation procedure. A non-uniform triangular tessellation adapted to the motion content of the partitions is employed.; A new efficient block-matching search algorithm that is equivalent to the full search block-matching algorithm is developed. It achieves high performance with low power dissipation. The proposed algorithm reduces the computational load of the computationally intensive full search with successive elimination of non-candidate blocks in the search window. This computational reduction leads to low-power VLSI implementation of the algorithm. Also, it leads to faster efficient motion estimation procedure. Simulation study shows that the new algorithm reduces the average search points by 88% compared to the full search block-matching algorithm and by 80% compared to other successive-elimination based block-matching algorithms. This algorithm is utilized in the two-stage global motion estimation technique.; Simulation results show that the proposed two-stage motion estimation technique performs better than the conventional ones in terms of the peak signal-to-noise ration (PSNR) and the compression ratio (CR).