Mixed Model Cross-Component Prediction Mode

Abstract

This invention provides a video encoding / decoding method and apparatus. [Solution] The method determines a feature value based on at least one of the reconfigured chroma samples near the current chroma block and the reconfigured luma samples near a luma block located at the same position as the current chroma block, and groups the chroma samples of the current chroma block and the luma samples of the luma block located at the same position as the current chroma block into multiple groups based on the threshold. Each of the multiple groups contains its own chroma sample and its own luma sample. The method also determines the cross-component prediction mode for each of the multiple groups by comparing each of the chroma samples and luma samples of each group with the determined feature value, and reconfigures the current chroma block based on this.

Images