Seunggeun Chi

Papers

Total Citations

H-Index

About

Seunggeun Chi is a researcher advancing the frontiers of computer vision, with a particular focus on human-centric understanding. His work primarily targets two challenging domains: skeleton-based action recognition and human pose estimation. Chi’s major contribution lies in tackling the critical problem of real-time performance and robustness to occlusion. His most notable work, "InfoGCN++," directly addresses a fundamental limitation of prior models by enabling online skeleton-based action recognition—allowing systems to classify actions without needing to observe the entire sequence. This innovation, which has already garnered 18 citations since its 2024 publication, is crucial for practical applications like human-robot interaction and autonomous driving. In a complementary vein, his paper "Pose Relation Transformer Refine Occlusions for Human Pose Estimation" (6 citations) adapts powerful NLP techniques to the visual domain, demonstrating how transformer architectures can effectively "fill in" missing or occluded body joints. By bridging the gap between state-of-the-art models and real-world constraints, Chi’s work is paving the way for more reliable and responsive AI systems that can perceive and anticipate human motion in dynamic environments.