Changhyun Choi

About

Changhyun Choi is a prominent robotics researcher whose work spans computer vision, robotic manipulation, and machine learning, with particular expertise in 3D perception, object grasping, and autonomous robotic systems. His early contributions established foundational techniques in real-time pose estimation and tracking, most notably his voting-based pose estimation algorithm (2012, 192 citations), which leveraged emerging 3D sensor technology to advance robotic assembly and computer vision applications. Building on this foundation, Choi has made significant strides in deep learning-driven grasping, demonstrating how 3D convolutional neural networks can enable soft robotic hands to manipulate unknown objects (2018, 172 citations) and pioneering the concept of "grasping the invisible," where robots intelligently combine pushing and grasping to locate and retrieve occluded targets (2020, 115 citations). His research extends into agricultural robotics, multi-robot team coordination, and human-robot interaction through virtual reality teleoperation, reflecting an impressively broad research vision. More recently, Choi has explored attribute-based grasping and collision-aware manipulation in constrained environments, continually pushing the boundaries of dexterous robotic autonomy. With hundreds of citations across a diverse body of work, Choi stands as an influential voice shaping the future of intelligent robotic systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Voting-based pose estimation for robotic assembly using a 3D sensor

   192 citations · 2012
2. 2
   Learning Object Grasping for Soft Robot Hands

   172 citations · 2018
3. 3
   A Deep Learning Approach to Grasping the Invisible

   115 citations · 2020
4. 4
   Stereo-vision-based crop height estimation for agricultural robots

   103 citations · 2020
5. 5
   Real-time 3D model-based tracking using edge and keypoint features for robotic manipulation

   95 citations · 2010
6. 6
   Multi-scale assembly with robot teams

   44 citations · 2015
7. 7
   Helping Robots Learn: A Human-Robot Master-Apprentice Model Using Demonstrations via Virtual Reality Teleoperation

   43 citations · 2020
8. 8
   Collision-Aware Target-Driven Object Grasping in Constrained Environments

   36 citations · 2021
9. 9
   Attribute-Based Robotic Grasping With Data-Efficient Adaptation

   23 citations · 2024
10. 10
    Learning Visual Affordances with Target-Orientated Deep Q-Network to Grasp Objects by Harnessing Environmental Fixtures

    22 citations · 2021

Key Collaborators

Contact & Links