Video Motion Refinement Using Reference Motion Information

The patent applies preliminary action by performing motion refinement using first motion information before final motion compensation. The method refines second motion information by incorporating first motion information obtained from a different reference picture, improving motion estimation accuracy before the actual coding process. This preliminary refinement step enhances video coding efficiency without requiring complex real-time processing during encoding.

The patent uses first motion information as an intermediary element to refine second motion information. By introducing this intermediate motion data from a different reference picture, the system bridges the gap between simple motion compensation and complex motion estimation, achieving improved coding efficiency through a moderate level of processing complexity.

The motion refinement process is performed in advance during the encoding phase, preparing refined motion information that can be reused. This preliminary action improves motion information accuracy for subsequent coding operations without causing time loss during real-time playback or decoding, as the refinement work is completed beforehand.

The patent copies motion information from a different reference picture (first motion information) to refine the motion information for the current picture. This copying approach leverages existing motion data to improve accuracy without requiring extensive new measurements or calculations, thereby reducing processing time while maintaining high precision.