File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Coarse-to-Fine Domain Adaptive Semantic Segmentation with Photometric Alignment and Category-Center Regularization
Title | Coarse-to-Fine Domain Adaptive Semantic Segmentation with Photometric Alignment and Category-Center Regularization |
---|---|
Authors | |
Issue Date | 2021 |
Citation | Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Virtual Conference, 19-25 June 2021, p. 4051-4060 How to Cite? |
Abstract | Unsupervised domain adaptation (UDA) in semantic segmentation is a fundamental yet promising task relieving the need for laborious annotation works. However, the domain shifts/discrepancies problem in this task compromise the final segmentation performance. Based on our observation, the main causes of the domain shifts are differences in imaging conditions, called image-level domain shifts, and differences in object category configurations called category-level domain shifts. In this paper, we propose a novel UDA pipeline that unifies image-level alignment and category-level feature distribution regularization in a coarse-to-fine manner. Specifically, on the coarse side, we propose a photometric alignment module that aligns an image in the source domain with a reference image from the target domain using a set of image-level operators; on the fine side, we propose a category-oriented triplet loss that imposes a soft constraint to regularize category centers in the source domain and a self-supervised consistency regularization method in the target domain. Experimental results show that our proposed pipeline improves the generalization capability of the final segmentation model and significantly outperforms all previous state-of-the-arts. |
Description | Paper Session Three: Paper ID 2873 |
Persistent Identifier | http://hdl.handle.net/10722/301300 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Ma, H | - |
dc.contributor.author | Lin, X | - |
dc.contributor.author | Wu, Z | - |
dc.contributor.author | Yu, Y | - |
dc.date.accessioned | 2021-07-27T08:09:05Z | - |
dc.date.available | 2021-07-27T08:09:05Z | - |
dc.date.issued | 2021 | - |
dc.identifier.citation | Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Virtual Conference, 19-25 June 2021, p. 4051-4060 | - |
dc.identifier.uri | http://hdl.handle.net/10722/301300 | - |
dc.description | Paper Session Three: Paper ID 2873 | - |
dc.description.abstract | Unsupervised domain adaptation (UDA) in semantic segmentation is a fundamental yet promising task relieving the need for laborious annotation works. However, the domain shifts/discrepancies problem in this task compromise the final segmentation performance. Based on our observation, the main causes of the domain shifts are differences in imaging conditions, called image-level domain shifts, and differences in object category configurations called category-level domain shifts. In this paper, we propose a novel UDA pipeline that unifies image-level alignment and category-level feature distribution regularization in a coarse-to-fine manner. Specifically, on the coarse side, we propose a photometric alignment module that aligns an image in the source domain with a reference image from the target domain using a set of image-level operators; on the fine side, we propose a category-oriented triplet loss that imposes a soft constraint to regularize category centers in the source domain and a self-supervised consistency regularization method in the target domain. Experimental results show that our proposed pipeline improves the generalization capability of the final segmentation model and significantly outperforms all previous state-of-the-arts. | - |
dc.language | eng | - |
dc.relation.ispartof | IEEE Conference on Computer Vision and Pattern Recognition (CVPR) | - |
dc.title | Coarse-to-Fine Domain Adaptive Semantic Segmentation with Photometric Alignment and Category-Center Regularization | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Yu, Y: yzyu@cs.hku.hk | - |
dc.identifier.authority | Yu, Y=rp01415 | - |
dc.identifier.hkuros | 323543 | - |
dc.identifier.spage | 4051 | - |
dc.identifier.epage | 4060 | - |