Abstract

This paper describes using independent context models for encoding and decoding end of blocks (EOBs) in video compression. Conventional approaches treat the chroma U and V components with a shared context model, assuming similar statistical distributions. However, empirical evidence shows that the U and V chroma channels have distinct statistical properties, especially when cross-component coding tools are used or enabled. For instance, transform coefficients in the V channel tend to be more concentrated in the low-frequency region compared to the U channel. Therefore, using separate context models for the chroma U and V components enables the codec to adapt more effectively to their specific distributions. This differentiation enhances compression efficiency without increasing the complexity of the encoder or decoder.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS