JunukCha
Junuk Cha
I am a graduate student in a combined Master's and PhD program specializing in Computer Vision and Deep Learning.

Junuk Cha

About Me

My name is Junuk Cha, and I am currently a combined MS/PhD student affiliated to the Artificial Intelligence Graduate School (AIGS) at UNIST since March 2021. I obtained BS (2021) degree from Dept. of Mechanical Engineering at UNIST.

Email: jucha@unist.ac.kr (official), junukcha@gmail.com (personal)

Lab link: UVL Lab webpage

Education

  • M.S./Ph.D. in Artificial Intelligence, UNIST, South Korea, Mar. 2021 - Feb. 2025 (expected)
  • B.S. in Mechanical Engineering, UNIST, South Korea, Mar. 2015 - Feb. 2021

Research Interests

  • Computer Vision and deep learning.
  • 3D human body pose estimation, 3D human body mesh reconstruction, 3D motion generation.
  • Action recognition and clothed human mesh reconstruction.
  • Text-guided 2D image editing.
  • Audio-driven talking head synthesis.

Researches

International Conference

  • 3DGS-based Bimanual Category-agnostic Interaction Reconstruction
    Jeongwan On, Kyeonghwan Gwak, Gunyoung Kang, Hyein Hwang, Soohyun Hwang, Junuk Cha, Jaewook Han, Seungryul Baek
    in Technical Report of European Conference on Computer Vision (ECCV), Milan, Italy, 2024.
    We got 1st place at the bimanual category-agnostic interaction reconstruction challenge.
    [Paper]
3DGS-based Bimanual Category-agnostic Interaction Reconstruction
  • Text2HOI: Text-guided 3D Motion Generation for Hand-Object Interaction
    Junuk Cha Jihyeon Kim, Jae Shin Yoon*, Seungryul Baek.*
    in Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), Seattle, USA, 2024.
    Co-last authors*.
    [Paper][Supp][Arxiv][Project]
text-guided 3D motion generation for hand-object interaction
  • VLM-PL: Advanced Pseudo Labeling Approach for Class Incremental Object Detection via Vision-Language Model
    Junsu Kim, Yunhoe Ku*, Jihyeon Kim*, Junuk Cha, Seungryul Baek.
    in Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR) workshop, Seattle, USA, 2024.
    Equal contribution*.
    [Paper][Arxiv]
Advanced Pseudo Labeling Approach for Class Incremental Object Detection via Vision-Language Model
  • 3D Reconstruction of Interacting Multi-Person in Clothing from a Single Image
    Junuk Cha, Hansol Lee, Jaewon Kim, Bao Truong, Jae Shin Yoon*, Seungryul Baek*.
    in Proc. of IEEE Winter Conf. on Applications of Computer Vision (WACV), Hawaii, USA, 2024.
    Co-last authors*.
    [PDF]
clothed mesh reconsturction of multi-person
  • Multi-Person 3D Pose and Shape Estimation via Inverse Kinematics and Refinement
    Junuk Cha, Muhammad Saqlain, GeonU Kim, Mingyu Shin, Seungryul Baek.
    in Proc. of European Conference on Computer Vision (ECCV), Tel Aviv, Israel, 2022.
    [PDF][Code]
mesh reconsturction of multi-person

International journal and arxiv

  • Dynamic Appearance Modeling of Clothed 3D Human Avatars using a Single Camera
    Hansol Lee, Junuk Cha, Yunhoe Ku, Jae Shin Yoon*, Seungryul Baek*.
    in arXiv 2023.
    Co-last authors*.
    [PDF]
reconstruction of dynamic appearance
  • HOReeNet: 3D-aware Hand-Object Grasping Reenactment
    Changhwa Lee, Junuk Cha, Hansol Lee, Seongyeong Lee, Donguk Kim, Seungryul Baek.
    in arXiv 2022.
    [PDF]
generation of hand-object image
  • Learning 3D Skeletal Representation From Transformer for Action Recognition
    Junuk Cha, Muhammad Saqlain, Donguk Kim, Seungeun Lee, Seongyeong Lee, Seungryul Baek.
    in IEEE Access 2022 (IF: 3.367).
    [PDF]
recognition of action
  • 3DMesh-GAR: 3D human body mesh-based method for group activity recognition
    Muhammad Saqlain, Donguk Kim, Junuk Cha, Changhwa Lee, Seongyeong Lee, Seungryul Baek.
    in Sensors 2022 (IF:3.576).
    [PDF]
recognition of group action
  • Towards Single 2D Image-level Self-supervision for 3D Human Pose and Shape Estimation
    Junuk Cha, Muhammad Saqlain, Changhwa Lee, Seongyeong Lee, Seungeun Lee, Donguk Kim, Won-Hee Park, Seungryul Baek.
    in Journal of Applied Sciences 2021 (IF:2.679).
    [PDF][Code]
self-supervised mesh reconstruaction

Internships

adobe Research Intern at Inria, Nice, France (Sep. 2024 - Nov. 2024)

adobe Research Scientist/Engineer Intern at Adobe, San Jose, CA, USA (Feb. 2024 - May 2024)

Academic Activities

  • Program Committee (Reviewer) at AAAI 2025.
  • Program Committee (Reviewer) at NIPS 2024.
  • Program Committee (Reviewer) at BMVC 2024.
  • Program Committee (Reviewer) at AAAI 2024.
  • Program Committee (Emergency-Reviewer) at BMVC 2023.
  • Program Committee (Reviewer) at AAAI 2023.

Patents

  • Dangerous Region Approaching Worker Detection and Stopping Machine Device Based on Thermal Images, 2023. Applicant : UNIST and Hyundai

Projects

  • Development of vision recognition module for ARTG control, Ministry of Oceans and Fisheries, Jun. 2023 ~ Dec. 2023
    • Detecting object and predicting multi agent’s future trajectory with velocity
  • Development of an optimized human detection system algorithm for production factory environments, Hyundai, Jun. 2022 ~ Dec. 2023
    • Detecting human segmentation masks and bounding boxes
    • Improving performance for corner cases such as occluded people

Teaching Assistant & Mentoring

  • Mentoring
    • 2nd Ulju AI 4.0 Studio at UNIST, Ulsan, South Korea, Mar. 2022 ~ Nov. 2022.
  • TA
    • Navy AI professionals training process TA, Jinhae, South Korea, 12 Sep. 2023.
    • 5th AI Novatus Academia PBL (Computer Vision part) at UNIST AI Innovation Park, Ulsan, South Korea, Mar. 2023 ~ Jun.2023.
    • 1st Gyungnam AI Novatus Academia (Computer Vision part) at UNIST AI Innovation Park, Masan, South Korea, 15 Apr. 2022.
    • 3rd AI Novatus Academia (Computer Vision part) at UNIST AI Innovation Park, Ulsan, South Korea, 25 Feb. 2022.
  • Lecture TA
    • 2023 Fall AI51601 Computer Vision.
    • 2021 Fall ITP11701 Introduction to AI Programming 2.