Constrained Merge Flag Signaling in Video Coding

The patent extracts and removes redundant mode syntax signaling from the bitstream. When only one mode is valid for a block (determined by block size constraints or SPS flags), the corresponding mode syntax is not signaled, as the mode can be inferred from the block characteristics. This eliminates unnecessary signaling while maintaining accurate mode indication.

The decoder uses the block characteristics (size, SPS flags) to self-determine which mode is valid without requiring explicit signaling. The block itself provides the information needed to infer the mode, eliminating the need for separate mode indication syntax when only one mode is possible.

The patent applies different signaling strategies to different blocks based on their local characteristics. For blocks where only one mode is valid (e.g., small blocks where subblock mode is not allowed), no mode syntax is signaled. For blocks where multiple modes are valid, the appropriate mode syntax is signaled. This localized approach optimizes both coding efficiency and signaling complexity.

Instead of always signaling all possible mode syntaxes (excessive action), the patent signals only the necessary mode syntax for each block (partial action). The amount of signaling is adjusted to match the actual mode selection needs, avoiding unnecessary flags while maintaining coding flexibility where required.