Picture Boundary Split Mode Selection in Multi-Tree Video Coding

The patent applies dynamics by making the split mode selection adaptive rather than fixed. The encoder dynamically chooses between implicit quad-split and explicit split modes based on the block's position relative to picture boundaries. This allows the system to transition from a rigid implicit split approach to a flexible approach that can adapt to different boundary scenarios, resolving the contradiction between signaling efficiency and partitioning flexibility.

The patent changes the parameter of split mode selection based on the block's position parameters. When a block is completely inside the picture, implicit quad-split is used; when it crosses the boundary, explicit split modes become available. This parameter-based adaptation allows the system to optimize signaling efficiency for interior blocks while gaining flexibility for boundary blocks, resolving the technical contradiction.

The patent applies local quality by treating blocks at different positions differently. Blocks completely inside the picture use implicit quad-split (local optimization for signaling efficiency), while blocks crossing boundaries use explicit split modes from a reduced set (local optimization for boundary handling). This localized approach allows the system to reduce overall bit budget while maintaining necessary flexibility at critical boundary locations.

The patent uses partial action by providing a reduced set of split modes rather than all possible split modes for boundary blocks. This partial provision is sufficient to handle boundary cases effectively while significantly reducing the bit budget compared to signaling all possible split modes, thus resolving the contradiction between bit budget reduction and maintaining adequate flexibility.

The patent makes the split derivation dynamic by conditioning it on the block's position relative to picture boundaries. The system transitions from static implicit derivation to dynamic selection between implicit and explicit modes, enabling adaptability to different boundary positions while maintaining signaling efficiency for cases where implicit derivation remains valid.

The patent changes the derivation parameter based on position parameters. The decision to use implicit or explicit split modes is determined by parameters describing the block's relationship to picture boundaries. This parameter-based adaptation allows the system to maintain signaling efficiency for interior blocks while gaining adaptability for boundary blocks with different position characteristics.