CIIP Video Prediction for Higher Coding Efficiency

The patent segments the prediction process into multiple candidate modes (inter prediction modes and intra prediction modes) and evaluates them separately before combining the best results. This segmentation allows the system to achieve higher coding efficiency by selecting from diverse prediction strategies without requiring a single complex prediction algorithm, thus resolving the contradiction between coding efficiency and device complexity.

The patent implements dynamic mode selection where the prediction mode is adaptively chosen based on the characteristics of each video block. The system dynamically evaluates multiple prediction modes and selects the optimal combination for each block, enabling higher coding efficiency through adaptive processing rather than static conventional methods, while maintaining manageable complexity through standardized evaluation procedures.

The patent performs preliminary evaluation of multiple prediction modes before final selection. By pre-calculating and comparing candidate prediction modes (both inter and intra) and their associated cost functions, the system prepares optimal prediction candidates in advance, achieving high coding efficiency without requiring complex real-time processing during encoding, thus resolving the contradiction between coding efficiency and processing complexity.

The patent implements a self-service mechanism where the encoding system automatically evaluates and selects the optimal prediction mode combination without external intervention. The standardized evaluation process autonomously compares multiple prediction modes and selects the best fit for each video block, achieving high coding efficiency through automated intelligent selection while keeping processing complexity manageable through systematic evaluation procedures.