AVC Video Stream Picture Interdependency Tier Extraction

The patent segments the complex picture interdependency structure into hierarchical layers (I-slices, P-slices, B-slices, and sub-slices). By organizing pictures into these segmented groups with defined reference relationships, the complex interdependencies are broken down into manageable units, allowing efficient random access and stream manipulation while maintaining AVC compression efficiency.

The patent introduces a hierarchical dimension to picture organization by creating multiple levels of slice structures (slices within groups, sub-slices within slices). This dimensional organization transforms the flat complex interdependency into a structured hierarchy, enabling more efficient processing and manipulation of video streams.

By segmenting the video stream into hierarchically organized slices and groups with explicit reference relationships, the patent enables easier stream manipulation operations such as random access, seeking, and extraction. Each slice group can be independently processed, making stream manipulation more manageable despite using advanced compression tools.

The patent performs preliminary organization of pictures into hierarchical slice structures during encoding. This pre-organization establishes clear reference relationships and access points, making subsequent stream manipulation operations easier and more efficient without requiring complex real-time analysis of picture dependencies.

The patent segments temporal redundancy exploitation into manageable slice groups with defined reference patterns. By organizing B-slices to reference specific I-slices and P-slices in a hierarchical manner, the complex temporal predictions are divided into structured units, reducing processing complexity while maintaining compression efficiency.

The patent applies different reference picture strategies to different slice groups locally. Each slice group can have its own reference picture selection and prediction mode, allowing optimized exploitation of temporal redundancies in different regions of the video stream without requiring complex global processing.