1 paper
Tiancheng Si, Wenlong Xu, Dahua Gao
Existing context models for learned image com- pression rely primarily on internal priors from the current image, overlooking common patterns and typical textures shared across nat…