Video Coding Template Matching Search Area Segmentation

The patent divides the search area into multiple regions (first search area, second search area, third search area) with different sizes and priorities. The first search area has a larger size and is searched first, while subsequent search areas are smaller and searched in sequence. This segmentation allows the system to explore a large overall search space without performing exhaustive searches across the entire area, thus improving prediction block identification while controlling computational complexity.

The patent performs preliminary actions by first determining and searching the first search area before proceeding to smaller search areas. The search process is structured to evaluate templates in the first search area initially, and only if needed, expand to subsequent search areas. This preliminary action approach allows the system to quickly identify good prediction blocks in the most likely regions without immediately committing computational resources to exhaustive searches.

The patent segments the search process into multiple phases corresponding to different search areas. The first search area is processed first with higher priority, and subsequent search areas are processed only if needed. This segmentation enables the system to achieve good coding performance by thoroughly searching the most promising regions while avoiding the time cost of exhaustively searching the entire large search area.

The patent implements a periodic or staged search approach where templates are evaluated in the first search area, and if satisfactory prediction blocks are not found, the search periodically expands to the second and third search areas. This periodic expansion allows the system to maintain efficiency by limiting detailed searches to necessary regions while still achieving improved coding performance when needed.