Merge Candidate List Ordering for Parallel Motion Compensation

The patent divides the neighboring blocks into specific sets (first neighboring blocks and second neighboring blocks) based on their positions relative to the current block. By segmenting the availability checking process into distinct groups with different processing requirements, the system can optimize prediction accuracy for critical blocks while reducing processing time for less critical ones, thus resolving the contradiction between prediction accuracy and processing time.

The patent applies different availability determination rules to different neighboring blocks based on their local positions. First neighboring blocks (with specific spatial relationships) are checked with stricter criteria to ensure high prediction accuracy, while second neighboring blocks use relaxed criteria to reduce processing time. This local differentiation allows the system to balance between prediction accuracy and processing time efficiently.

The patent introduces dynamic processing modes for motion candidate list configuration. The system can operate in sequential mode for simplicity or switch to parallel processing mode for enhanced efficiency. By making the processing approach dynamic and adaptable, the system maintains simplicity when needed while achieving high encoding/decoding efficiency when parallel processing is enabled, thus resolving the contradiction between processing simplicity and efficiency.

The patent transitions from one-dimensional sequential processing to multi-dimensional parallel processing by introducing separate processing threads or stages for different neighboring block sets. This dimensional change in processing architecture enables simultaneous operation on different blocks, significantly improving encoding/decoding efficiency while maintaining manageable system complexity through structured parallelization.

The patent applies different availability determination rules to different neighboring blocks based on their local positions. First neighboring blocks (with specific spatial relationships) are checked with stricter criteria to ensure high prediction accuracy, while second neighboring blocks use relaxed criteria to reduce processing time. This local differentiation allows the system to balance between prediction accuracy and processing time efficiently.

The patent changes the availability determination parameters based on the type and position of neighboring blocks. Different parameters and thresholds are applied to first versus second neighboring blocks, allowing the algorithm to adapt its behavior to local conditions. This parameter differentiation improves prediction performance by tailoring the availability criteria to the specific characteristics of each neighboring block region.