Hardware Entropy Engine Count Table Generation for Video Coding Adaptation

The patent segments the probability table adaptation process into two parts: a count table that stores only the necessary statistical information (segmented from the full probability table), and the probability table itself. This segmentation allows selective storage of only the count information needed for adaptation, reducing overall storage requirements while maintaining encoding/decoding performance.

The patent extracts only the essential count information from the complete probability table structure. Instead of storing and processing the entire probability table for adaptation purposes, only the count table containing statistical counts is extracted and stored in external memory, reducing storage burden while enabling effective probability table adaptation.

The patent introduces a count table as an intermediary structure between the video data and the probability table. This count table serves as a mediator that stores pre-computed statistical information, simplifying the adaptation process by providing ready-to-use count data rather than requiring complex real-time statistical analysis during encoding/decoding operations.

The patent performs preliminary computation of count statistics during the encoding/decoding process and stores them in the count table for later use. By pre-computing and storing these statistical counts, the system avoids the need for complex real-time calculations during probability table adaptation, thereby reducing processing complexity.