Yongjun Choi

M.S. Student at UNIST
3D Vision & Robotics Lab

3D Scene Understanding Generative Model Video Understanding Vision-Language Model Audio-Visual Modeling

Hello! I am a master's student in the 3D Vision & Robotics Lab at UNIST, advised by Prof. Kyungdon Joo.

I was a visiting student in the CARTE, MIE, at the University of Toronto through the AI Convergence Program, supported by the IITP, Korean Government.

My research has focused on 3D scene understanding, image manipulation using generative model, video understanding, and visual–language modeling. Currently, I'm expanding my research scope to encompass audio-visual modeling, where I am actively developing and evaluating multimodal learning frameworks.

Contact: ccyjun123@unist.ac.kr

CV LinkedIn GitHub

Publications

AnyBald: Toward Realistic Diffusion-Based Hair Removal In-the-Wild Yongjun Choi, Seungoh Han, Soomin Kim, Sumin Son, Mohsen Rohani, Edgar Maucourant, Dongbo Min, Kyungdon Joo ACCEPTED WACV 2026 *Equal contribution [Project Page] Developed AnyBald, a mask-free diffusion-based framework for realistic hair removal in the wild. Achieves natural bald manipulation while preserving facial identity, enabling robust performance across diverse real-world scenarios.
RAC-VAD: Reference-Guided Temporal Alignment and Pairwise Comparison for Video Anomaly Detection in Display Inspection Yongjun Choi, Gyeongsu Cho, Jinhyeok Kim, Changsu Ha, Sanggyu Biern, Kyungdon Joo UNDER REVIEW IEEE TCSVT Video-based anomaly detection system for industrial display device inspection, capable of identifying defects during long-term multi-device operation. Introduce two-stage based video alignment and comparison approach for robust, noise-resistant anomaly detection.
Demonstrating a Vision-Based AI Robot for Strategic Board Games Taehwan Kim, Dokeun Lee, Seonghyeon Kim, Yongjun Choi, Sungjun Heo, Thi Thuy Ngan Duong, Kyungdon Joo, Namhun Kim, Jeong hwan Jeon, Hyemin Ahn Technical Report *Equal contribution [Project Page] Developed a low-cost human–robot interaction system for playing Gomoku, integrating real-time vision perception, RL-based decision-making, and robotic arm control. Offers an affordable, real-world gaming experience that merges physical gameplay with intelligent decision-making.

Selected Projects

Realistic Hair Removal and Reconstruction in Images 2025 Research project (Working with Modiface) Developed a robust hair removal framework primarily designed for hair transfer applications, enabling realistic and consistent bald rendering across diverse scenarios. Served as the technical lead in implementing the core diffusion-based model and proposing the initial data augmentation pipeline for robust training.
Detecting Anomalies from normal videos 2024 Industrial project (funded by Samsung Electronics) Developed video anomaly detection system for industrial display inspection, built to ensure stable defect detection during long-term multi-device operation. Served as the technical lead, spearheading the implementation of the core proposed method and conducting all major experiments.
Gomoku AI: Demonstrating a Vision-Based AI Robot for Strategic Board Games 2024 HRI course final project Built a vision-based human–robot interaction system for playing Gomoku, integrating real-time perception, AI decision-making, and robotic arm control. Responsible for the entire vision module, developing the perception system that enables the robot to recognize and understand the full game state.
Lang-Grouping: Object-centric semantic grouping for better understanding 3D scenes 2024 3D vision course final project · Apr – Jun 2024 Object-level language–3D scene understanding framework, reducing inference time compared to LangSplat. Propose an object-centric contrastive learning approach to enhance multi-view consistency for CLIP feature distillation in Gaussian Splatting.

Experience

Visiting Student

University of Toronto

Jan 2025 – Jun 2025

Special MEng student at MIE – completed 4 graduate-level courses with a GPA of 3.95/4.0
Industrial project with Modiface (Technical Leading)

Teaching Assistant

UNIST

Jul 2024 – Dec 2024

AI Programming 2 (Sep 2024 – Dec 2024)
Kyungnam Novatus Academia (Jul 2024)

Research Assistant

3D Vision & Robotics Lab, UNIST

Mar 2024 – Aug 2026

Video understanding, Language-guided 3D scene understanding
Industrial project with Samsung Electronics (Technical Leading)

Software Developer Intern

Upsight

Jun 2023 – Aug 2023

Contributed to developing a building crack detection model integrated into diagnostic processes
Contributed to the initial development of a landlord–tenant community app using Flutter

Undergraduate Research Intern

Computer Vision Lab, University of Seoul

Apr 2022 – Jul 2023

Mainly studied Masked Image Modeling (MIM)-based ViT and CAM methods
Built AI-based plant feature extractor for urban forest management mobile platform

Awards & Honors

🏆 3rd place, AWS DeepRacer Competition, BigData Winter Camp (2022)
🏆 3rd place, Engineering Mathematics Competition, University of Seoul
Awarded for three consecutive years: 2021, 2022, 2023
🏆 3rd place, Spatial Convergence Big Data Idea Competition (2023)
Proposed and developed the company's core concept as part of the team
🏆 3rd place, Syncathon Season 3 (AI development competition) — "finSET" (2023)
Serving as the team leader and project coordinator

Scholarships

💰 AI Convergence Program Scholarship, IITP, Korean Government (2025)
💰 Wonhae Foundation (2023)
💰 Boeing Korea (2022)
💰 Kochon Foundation (2022)

Education

UNIST — M.S., Artificial Intelligence Graduate School (Expected Aug 2026), GPA: 4.03/4.3
Advisor: Prof. Kyungdon Joo
University of Seoul — B.S., Electrical & Computer Engineering (Feb 2024), GPA: 3.98/4.5 (Major 4.06/4.5)
Advisor: Prof. Yongchul Kim