Papers
4
Total Citations
541
H-Index
3
About
Seungryul Baek is a computer vision researcher specializing in 3D hand pose estimation, egocentric action recognition, and human-computer interaction. His work sits at the intersection of RGB-D video analysis and deep learning, with a particular focus on understanding how hands interact with objects from a first-person perspective — a domain critical to advancing robotics, augmented reality, and virtual reality applications. Baek is perhaps best known for creating the **First-Person Hand Action Benchmark**, a large-scale dataset comprising over 100,000 RGB-D frames spanning 45 daily hand action categories and 26 distinct objects. This resource has become a cornerstone reference in the field, accumulating over 500 citations and providing the community with a rigorous foundation for training and evaluating hand action recognition systems. His continued contributions to benchmarking are evident in his 2024 work on pose estimation challenges for egocentric hand-object interactions, reflecting his sustained commitment to pushing the boundaries of holistic 3D scene understanding. Through carefully constructed datasets and evaluation frameworks, Baek has helped define standards for how researchers measure progress in egocentric vision — making his work especially valuable for students and practitioners building the next generation of immersive and intelligent systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4