Kyungwon Cho

MS/PhD Student @ Seoul National University

I am a first-year MS/PhD student at SNU Visual Computing Lab, advised by Prof. Hanbyul Joo. I'm interested in computer vision and robotics. My research focuses on egocentric vision, human motion, human-robot interaction, and world models.

Kyungwon Cho

News

Research

  1. NeurIPS 2026

    OmniEgoCap: Device-Agnostic Sequence-Level Egocentric Motion Reconstruction

    Kyungwon Cho, Jeonghyeon Na, Hanbyul Joo

    Project Page · arXiv · Code · OmniEgoDB

    OmniEgoCap is a sequence-level diffusion framework that leverages long-range physical invariants and geometry-aware augmentation to reconstruct consistent 3D full-body motion across diverse egocentric camera setups.

  2. arXiv 2026

    OmniRobotHome: A Multi-Camera Home Platform for Real-Time Human-Robot Interaction

    Junyoung Lee*, Inhee Lee*, Kyungwon Cho*, Jeonghwan Kim, Mingi Choi, Sookwan Han, Lee Chae-Yeon, Wonjung Woo, Gunhee Kim, Jisoo Kim, Jeonghyeon Na, Hanbyul Joo
    * Equal contribution

    Project Page · arXiv

    OmniRobotHome is a multi-camera home platform with 48 hardware-synchronized cameras and three manipulators that provides real-time markerless full-body human pose, 6D object pose, and anticipatory motion forecasting for safe and assistive human-robot interaction.

  3. arXiv 2026

    DeVI: Physics-based Dexterous Human-Object Interaction via Synthetic Video Imitation

    Hyeonwoo Kim, Jeonghwan Kim, Kyungwon Cho, Junyoung Lee, Hanbyul Joo

    Project Page · arXiv · Code

    DeVI is a physics-based framework that imitates text-conditioned synthetic videos to learn dexterous human-object interaction, using a hybrid 3D-human and 2D-object tracking reward to achieve physically realistic, zero-shot generalizable manipulation.

Awards