1 paper
Gengyu Liao, Yechun Yong, Zheng Chen +4
Open-domain video super-resolution (VSR) requires temporally coherent details across widely varying scenes and subjects. However, existing training datasets are often semantically…