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...
Главные авторы: | Zhang, F, Torr, P, Ranftl, R, Richter, S |
---|---|
Формат: | Conference item |
Язык: | English |
Опубликовано: |
Neural Information Processing Systems Foundation
2021
|
Схожие документы
-
Open vocabulary semantic segmentation with Patch Aligned Contrastive Learning
по: Mukhoti, J, и др.
Опубликовано: (2023) -
Scalable cascade inference for semantic image segmentation
по: Sturgess, P, и др.
Опубликовано: (2012) -
Dense semantic image segmentation with objects and attributes
по: Zheng, S, и др.
Опубликовано: (2014) -
Pyramid Context Contrast for Semantic Segmentation
по: Yuzhong Chen, и др.
Опубликовано: (2019-01-01) -
Highly Contrast Image Correction for Dim Boundary Separation of Image Semantic Segmentation
по: Jinyeob Choi, и др.
Опубликовано: (2021-01-01)