1 paper · 2 filters
Hao Yan, Zhihui Ke, Xiaobo Zhou +3
Implicit neural representations for video (NeRV) have recently become a novel way for high-quality video representation. However, existing works employ a single network to represen…