Intra Prediction Mode Encoding Using Candidate Sets

The patent divides the intra-prediction modes into multiple candidate mode sets (first candidate mode set, second candidate mode set, third candidate mode set) with different numbers of modes. Instead of transmitting all modes uniformly, the system segments the mode selection process into hierarchical stages, where each stage narrows down the candidate pool. This segmentation allows the encoder to signal only the necessary information to identify the selected mode within the appropriate candidate set, reducing the bit rate while maintaining access to multiple prediction modes for high accuracy.

The patent applies preliminary action by pre-defining multiple candidate mode sets with different characteristics before the actual encoding process. The first candidate mode set, second candidate mode set, and third candidate mode set are prepared in advance with specific mode compositions. During encoding, the system only needs to signal which candidate set is selected and the index within that set, rather than evaluating all possible modes from scratch. This preliminary organization of candidate modes simplifies the encoding decision process while achieving bit rate reduction.