Papers
2
Total Citations
11
H-Index
2
About
Sanghyun Woo is a leading researcher in computer vision, specializing in domain adaptation and semantic segmentation—critical areas for enabling autonomous systems to operate reliably in unpredictable, real-world environments. His work addresses the fundamental challenge of deploying deep recognition models in dynamic settings where training data cannot cover all possible scenarios. Woo’s most notable contribution is the **“Discover, Hallucinate, and Adapt”** framework for Open Compound Domain Adaptation (OCDA) in semantic segmentation, which introduced a novel method for handling multiple, unknown target domains simultaneously. This work, with 7 citations, has become a reference point for researchers tackling label-scarce applications like autonomous driving and medical imaging. Expanding on this, Woo developed a **Test-Time Adaptation (TTA)** approach with compound domain knowledge management, enabling models to continuously adapt during deployment—a breakthrough for lifelong robotic learning. With 4 citations, this work demonstrates his focus on practical, deployable AI. Woo’s research bridges the gap between static training and dynamic reality, making him a key figure in advancing robust, adaptive vision systems for real-world use.
Research Focus
Key Achievements
Top Papers
- 1
- 2