Historical Motion Vector Reset for First-Block Prediction

The patent divides the frame into two distinct sets of blocks: first blocks of each row and other blocks. Different historical motion vector predictor lists are maintained for these two sets. The first blocks use a dedicated list that is reset for each row, while other blocks use a shared historical list. This segmentation allows optimized prediction for first blocks without being constrained by the limited storage capacity of a single unified list.

The patent performs preliminary actions by resetting the historical motion vector predictor list specifically for first blocks of each row before processing those blocks. This reset operation prepares the list with appropriate initial values (such as zero vectors or vectors from previous rows) that are suitable for predicting motion in the first blocks, ensuring accurate prediction before the main processing begins.

The patent segments the coding process into two distinct pathways: one for first blocks of each row and another for other blocks. This segmentation enables wavefront parallel coding to proceed efficiently while maintaining prediction accuracy for first blocks through the dedicated reset list mechanism, resolving the conflict between processing efficiency and prediction accuracy.

The reset list acts as an intermediary structure between the wavefront parallel coding process and the motion prediction requirement. It provides the necessary historical motion vector data for first blocks without interfering with the parallel processing efficiency, mediating between the need for accurate prediction and the benefits of wavefront parallel coding.