← back
arXivXiang Fan, Yuheng Wang, Bohan Fang, Zhongzheng Ren, Ranjay KrishnaThu, May 14, 2026, 10:59 AM PDT
score 9.2

Better video generation by conditioning the decoder with reference images

Original: RefDecoder: Enhancing Visual Generation with Conditional Video Decoding

Source: arxiv.org

Writing ELI5 summary…