Adaptive Search Range Method for Motion Estimation

The search range is divided into multiple regions, and the algorithm selectively searches only the relevant regions where motion or disparity vectors are likely to be found, rather than searching the entire broad search range. This segmentation reduces computation time while maintaining coding quality by concentrating search efforts in areas of interest.

The algorithm performs motion and disparity estimation on only a subset of blocks (e.g., non-skipped blocks, or blocks with high prediction error) rather than all blocks in the frame. This partial action approach reduces overall computation time while maintaining sufficient coding quality for the majority of the video content.

The patent applies motion and disparity estimation selectively to only certain blocks (e.g., non-skipped blocks, blocks with high residual error, or blocks in regions of interest) rather than all blocks. This partial application of the estimation process significantly reduces computational resource consumption while maintaining adequate coding quality for the overall video stream.

Different blocks are treated differently based on their local characteristics. Blocks that require high precision (e.g., those with high prediction error or in important regions) receive full motion and disparity estimation, while other blocks use simpler or skipped modes. This local differentiation optimizes the balance between coding quality and computational resource usage.