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...
Glavni autori: | Zhang, F, Torr, P, Ranftl, R, Richter, S |
---|---|
Format: | Conference item |
Jezik: | English |
Izdano: |
Neural Information Processing Systems Foundation
2021
|
Slični predmeti
-
Open vocabulary semantic segmentation with Patch Aligned Contrastive Learning
od: Mukhoti, J, i dr.
Izdano: (2023) -
Scalable cascade inference for semantic image segmentation
od: Sturgess, P, i dr.
Izdano: (2012) -
Dense semantic image segmentation with objects and attributes
od: Zheng, S, i dr.
Izdano: (2014) -
Pyramid Context Contrast for Semantic Segmentation
od: Yuzhong Chen, i dr.
Izdano: (2019-01-01) -
Highly Contrast Image Correction for Dim Boundary Separation of Image Semantic Segmentation
od: Jinyeob Choi, i dr.
Izdano: (2021-01-01)