Taein Kwon

Papers

Total Citations

H-Index

About

Taein Kwon is a computer vision researcher whose work sits at the intersection of egocentric perception, human body understanding, and mixed reality. His research focuses on developing datasets and methods that enable machines to understand how people move, interact, and apply physical force — particularly from the first-person camera perspective afforded by head-mounted devices. Kwon's most impactful contribution, **EgoBody** (2022), addresses the challenging problem of estimating full human body shape and motion during social interactions using egocentric cameras, garnering 71 citations and establishing a foundational benchmark in the field. His work on **Context-Aware Sequence Alignment** (2022) advances temporal understanding of fine-grained human actions in video, with applications spanning robotics and mixed reality. Most recently, **EgoPressure** (2025) pushes the frontier further by tackling hand pressure and pose estimation from egocentric views — a uniquely difficult problem due to limited training data and occlusion challenges. Across these contributions, Kwon consistently addresses data scarcity by introducing richly annotated datasets alongside novel methodologies. His research has direct implications for augmented reality systems, human-robot interaction, and embodied AI, making him a noteworthy voice in the growing egocentric vision community.