Looking beyond single images for contrastive semantic segmentation learning

We present an approach to contrastive representation learning for semantic segmentation. Our approach leverages the representational power of existing feature extractors to find corresponding regions across images. These cross-image correspondences are used as auxiliary labels to guide the pixel-lev...

Celý popis

Podrobná bibliografie
Hlavní autoři: Zhang, F, Torr, P, Ranftl, R, Richter, S
Médium: Conference item
Jazyk:English
Vydáno: Neural Information Processing Systems Foundation 2021