Video Coding Bi-Prediction with Multi-Block Vectors for IBC

The patent divides the prediction process into multiple segments by using multiple block vectors (first BV, second BV, third BV) to represent different motion characteristics within the same block. Each block vector corresponds to a specific motion model (uni-directional, bi-directional, or tri-directional), allowing the system to segment the prediction task into manageable parts while maintaining high coding efficiency through selective application of appropriate motion models.

The patent implements dynamic selection of block vectors based on motion characteristics. The system dynamically determines which block vectors to use (one, two, or three) depending on the specific motion pattern detected in each block, allowing the complexity to adapt rather than remain fixed. This dynamic approach enables the system to use simple uni-directional prediction when appropriate while resorting to complex bi- or tri-directional prediction only when necessary.

The patent applies partial action by using multiple block vectors selectively rather than always. The system evaluates motion characteristics and applies only the necessary number of block vectors (one, two, or three) for each block, avoiding unnecessary computational overhead. This partial application of complex prediction methods only where needed maintains high coding performance while minimizing processing time compared to always using maximum complexity.

The patent changes the parameter of block vector quantity dynamically based on motion characteristics. By adjusting the number of block vectors from one to three depending on the motion pattern, the system optimizes the trade-off between processing time and coding performance. This parameter adaptation allows faster processing when simple motion models suffice while reserving more time for complex motion analysis only when necessary.