Summary of Open-set Domain Adaptation For Semantic Segmentation, by Seun-an Choe et al.
Open-Set Domain Adaptation for Semantic Segmentation
by Seun-An Choe, Ah-Hyung Shin, Keon-Hee Park, Jinwoo Choi, Gyeong-Moon Park
First submitted to arxiv on: 30 May 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed Open-Set Domain Adaptation for Semantic Segmentation (OSDA-SS) tackles the challenge of unsupervised domain adaptation in semantic segmentation by introducing unknown classes in the target domain. The key issues addressed are predicting the exact boundary and shape of these unknown classes. To achieve this, the authors propose Boundary and Unknown Shape-Aware open-set domain adaptation (BUS), which utilizes a novel dilation-erosion-based contrastive loss for discerning boundaries between known and unknown classes. Additionally, OpenReMix is introduced as a domain mixing augmentation method to guide the model in learning size-invariant features for shape detection. The proposed approach demonstrates significant improvements over previous methods in detecting unknown classes in the OSDA-SS scenario. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper solves a big problem in computer vision by allowing machines to learn from one task and apply it to another, even if some things are completely new. Currently, this is not possible because computers get confused when they see something they’ve never seen before. The authors developed a way for computers to tell apart what they know from what’s new, so they can make better predictions. This breakthrough could lead to machines being able to learn from lots of different tasks and apply them to real-world problems. |
Keywords
* Artificial intelligence * Contrastive loss * Domain adaptation * Semantic segmentation * Unsupervised
