Tile-Based Image Decoding for Parallel Processing and Synchronization

The image is divided into multiple independent tiles that can be processed in parallel. Each tile is self-contained with its own processing pipeline, allowing simultaneous decoding and rendering without requiring complex synchronization between tiles. This segmentation enables the system to achieve high processing efficiency while maintaining simple synchronization requirements.

Subordinate information such as subtitles and advertisements is segmented into separate tile regions within the image stream. This allows the information to be transmitted efficiently as part of the image data structure while enabling selective processing and display. The segmentation approach provides user convenience by integrating multiple information types without proportionally increasing bandwidth consumption.

The patent applies different motion prediction strategies to different tile regions based on local characteristics. Constrained tiles use motion prediction only within their own boundaries, while non-constrained tiles can reference other tiles for improved accuracy. This local quality approach optimizes decoding accuracy where needed while maintaining simpler processing in regions where it is sufficient, thereby balancing precision and complexity.