Subblock Intra-Prediction Partitioning for Image Encoding Order

The current block is divided into multiple subblocks, and intra-prediction is performed separately for each subblock. This segmentation allows the prediction to capture local variations in different regions of the block, thereby improving prediction accuracy while managing complexity through structured subdivision.

Different prediction modes are applied to different subblocks based on their local characteristics. Each subblock can select from multiple candidate prediction modes, allowing the system to adapt to local image features and improve overall prediction accuracy by treating each region according to its specific needs.

A candidate group of partition types is pre-configured and prepared before the actual prediction process. This preliminary configuration allows the decoder to efficiently select from predefined options without requiring transmission of detailed partitioning information, thereby improving efficiency while minimizing information loss through compact representation.

The system changes partitioning parameters by configuring multiple candidate partition types with different characteristics. By varying partitioning parameters and selecting the most appropriate candidate for each block, the system improves prediction efficiency while maintaining compact signaling to minimize information loss.