Chroma Adaptive Loop Filtering with Transpose-Based Block Classification

The chroma block is divided into multiple sub-blocks for independent classification and filtering. Each sub-block can be classified into different categories based on local characteristics, allowing targeted application of geometric transformations and filter coefficients. This segmentation enables precise processing of different chroma regions while maintaining manageable complexity through localized operations.

Different geometric transformations and filter coefficients are applied to different chroma blocks based on their local characteristics. The classification system identifies specific patterns in chroma blocks and applies appropriate processing parameters locally, rather than using a uniform approach across the entire image. This local quality principle improves video quality by adapting to local chroma variations while keeping the overall system complexity controlled through selective processing.

A classification system is established beforehand that categorizes chroma blocks based on their characteristics. The geometric transformations and filter coefficients are pre-determined for each class, allowing the encoder to quickly select appropriate processing parameters without performing complex real-time calculations. This preliminary classification approach improves coding efficiency by reducing the computational burden during actual filtering operations.

The system changes processing parameters (geometric transformations and filter coefficients) based on the classified characteristics of chroma blocks. By establishing parameter sets for different classification categories, the system can efficiently adapt to various chroma patterns without requiring complex computational resources. The parameter changes are predetermined and selected based on classification results, balancing coding efficiency with computational complexity.