2021
DOI: 10.48550/arxiv.2104.02488
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Weakly supervised segmentation with cross-modality equivariant constraints

Abstract: Weakly supervised learning has emerged as an appealing alternative to alleviate the need for large labeled datasets in semantic segmentation. Most current approaches exploit class activation maps (CAMs), which can be generated from image-level annotations. Nevertheless, resulting maps have been demonstrated to be highly discriminant, failing to serve as optimal proxy pixel-level labels. We present a novel learning strategy that leverages self-supervision in a multi-modal image scenario to significantly enhance… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
1
0

Year Published

2021
2021
2021
2021

Publication Types

Select...
1

Relationship

0
1

Authors

Journals

citations
Cited by 1 publication
(1 citation statement)
references
References 53 publications
(89 reference statements)
0
1
0
Order By: Relevance
“…Weakly-supervised Learning for Image Segmentation. Recent research has explored weak annotations to supervise models, including: bounding boxes [15], image-level labels [24], point clouds [25], and scribbles [17,4,8,32]. Although it is possible to extend the proposed approach to other types of weak annotations, herein, we focus on scribbles, which have shown to be convenient to collect in medical imaging, especially when annotating nested structures [4].…”
Section: Related Workmentioning
confidence: 99%
“…Weakly-supervised Learning for Image Segmentation. Recent research has explored weak annotations to supervise models, including: bounding boxes [15], image-level labels [24], point clouds [25], and scribbles [17,4,8,32]. Although it is possible to extend the proposed approach to other types of weak annotations, herein, we focus on scribbles, which have shown to be convenient to collect in medical imaging, especially when annotating nested structures [4].…”
Section: Related Workmentioning
confidence: 99%