SCLIP: Rethinking Self-Attention for Dense Vision-Language Inference.

Feng Wang, Jieru Mei, Alan L. Yuille

Browse the full ECCV paper archive.