An Efficient Three-Step Block Matching Motion Estimation Algorithm

To reduce the computational complexity of motion estimation algorithms and improve the reliability of sequential image super-resolution reconstruction, we propose an efficient three-step search algorithm. The algorithm implements a multi-step search strategy, beginning with coarse search operations. Leveraging the center-biased distribution characteristics of motion vectors and parallel processing concepts, it employs diamond-shaped small templates instead of traditional square templates for refined searching in optimal matching regions. This enhancement enables more accurate motion vector estimation through directional pattern optimization in code implementation, consequently improving image reconstruction quality. Experimental results confirm that the algorithm significantly reduces computational time while maintaining search accuracy, providing a more efficient and reliable solution for super-resolution image reconstruction. The core algorithm can be implemented using nested search loops with adaptive step size reduction and early termination mechanisms for computational optimization.