Motion Vector Candidate List Scanning With Limited Scaling

The patent applies partial action by limiting the number of scaling times to a specific value (e.g., 1 or 2 times) rather than performing exhaustive scaling operations. This selective approach scans candidate predicted motion vectors with controlled scaling iterations, achieving sufficient prediction accuracy while avoiding the exponential increase in computational complexity that would result from unlimited scaling operations.

By performing scaling operations only a limited number of times (partial action) rather than exhaustively scaling all candidate vectors multiple times, the patent achieves acceptable prediction accuracy while significantly reducing the total encoding time. This controlled approach prevents the time complexity from becoming prohibitive.

The patent changes the parameters of motion vector prediction by introducing controlled scaling operations with limited iterations. This modifies the prediction process to achieve better motion compensation accuracy for high-resolution images without proportionally increasing the data amount, as the scaling is applied selectively to candidate vectors rather than to the entire image data.