cv

General Information

Full Name Deok-Kyeong (DK) Jang
Date of Birth 14th October 1991
Nationality South Korea
Languages Korean (native), English (Limited working proficiency)

Research Interests

  • My research goal is to improve the quality of virtual character motion in computer graphics and AR/VR systems. I currently focus on the LiDAR based 3D motion capture/tracking, real-time motion stylization/characterization, and motion synthesis/control.

Education

  • 2017 - 2022
    Ph.D. in Computer Graphics / Motion Computing Lab - KAIST
    Korea Advanced Institute of Science and Technology, South Korea
    • Advisor: Prof. Sung-Hee Lee.
    • Research on
      • Motion stylization and characterziation.
      • Constructing motion manifold space.
      • Motion synthesis and control.
  • 2015 - 2016
    M.S. in Computer Graphics / Motion Computing Lab - KAIST
    Korea Advanced Institute of Science and Technology, South Korea
    • Advisor: Prof. Sung-Hee Lee.
    • Research on
      • Regresson-based landmark detection of human body.
  • 2009 - 2015
    B.S. in Physics and Mathematics - KAIST
    Korea Advanced Institute of Science and Technology, South Korea
    • Research on
      • Modeling the prey-predator system.

Work Experience

  • 2023 - NOW
    CTO / Co-founder at MOVIN (무빈 in Korean)
    • Product: 3D full-body motion capture solution using a single LiDAR.
    • Constructing a high-quality dataset featuring diverse subjects, containing synchronized LiDAR point cloud and optical motion capture data for a wide range of actions.
    • Developed real-time full-body motion capture framework based on a single LiDAR, incorporating global translation tracking.
  • 2023 - NOW
    Postdoctoral Researcher at Motion Computing Lab
    KAIST, Daejeon, South Korea
    • Advisor: Sung-Hee Lee.
    • Research on LiDAR based real-time 3d motion capture and motion synthesis.
  • 2022
    Research Scientist Intern at Meta Reality Lab
    Gemini team, Meta Reality Labs, Redmond, WA, USA
    • Manager: Dr.Yuting Ye, Research Scientist.
    • Collaborators:. Prof.Jungdam Won
    • Research on motion characterization in real-time, enhancement of motion style transfer and retargeting with various input sensors.

Projects

  • 2022 - 2023
    Motion tracking and characterization research for virtual avatars
    Meta Platforms Technologies
    • As a researcher of the project, developed real-time motion characterization framework for virtual avatars.
  • 2020 - 2022
    Study of styled motion generation for non-verbal communication of virtual human agents
    National Research Foundation of Korea
    • As a leading researcher of the project, developed humanoid agent’s appearance-style customizable motion generation framework.
  • 2017 - 2021
    Development of 4D reconstruction and dynamic deformable action model based hyper realistic service technology
    Ministry of Science, ICT and Future Planning, Giga Korea Project
    • As a leading researcher of the project, developed motion style transfer method and plugins to automatically generate stylized motion.
  • 2015 - 2017
    Development of simulation software for human body-sport gear complex for rapidly prototyping customized sports gear
    Ministry of Culture, Sports and Tourism
    • As a main developer of the project, developed an sports gear modeling technique that fits the personalized foot shape and sports gear interaction.

Paper Reviewer

  • SIGGRAPH, SIGGRAPH ASIA, Pacific Graphics, Computer Graphics Forum, IEEE Transactions on Visualization and Computer Graphics (TVCG)

Technical Skills

  • Programming Languages: Python, C#, C/C++, MATLAB
  • Operating Systems: Mac OS, Linux, Windows
  • Frameworks & Libraries: Pytorch, Tensorflow, Numpy, Eigen, igl, etc.
  • Tools: Unity, Blender, Docker system, etc.

Other Interests

  • Sports: Working out, Judo, Soccer, Swimming.
  • Hobbies: Traveling, Watching movies, Playing games.