Taehoon Kim

---------------------------------------------------------------------------------------------------------------------------------------------------

Education

Ajou University   Suwon, Korea

Ph.D in Department of Artificial Intelligence     Sep.  2022 - Present

Ajou University   Suwon, Korea

M.S in Department of Artificial Intelligence Sep.  2020 - Aug. 2022 

Ajou University   Suwon, Korea

B.S in Software and Computer Engineering   Mar. 2014 - Aug.2020

---------------------------------------------------------------------------------------------------------------------------------------------------

REsEARCH EXPERIENCEs

Ajou University   Suwon, Korea

Research Assistant, Computer Vision Lab             Sep. 2020 - Present  

University of Birmingham,           Birmingham, UK

Visiting Academics, Intelligent Robotics Lab Joint / Research Topic: Unsupervised Domain Adaptation                                         Sep. 2021 - Aug.2022

Mentor: Prof. Hyung Jin Chang              

Ajou University   Suwon, Korea

Research Intern, Computer Vision Lab   Sep. 2019 - Aug. 2020 

---------------------------------------------------------------------------------------------------------------------------------------------------

Publications

INTERNATIONAL CONFERENCES & Journal

• Seokhwa Cheung, Seungbeom Woo, Taehoon Kim, Wonjun Hwang, Knowledge tailoring: Bridging the teacher-student gap in semantic segmentation, Pattern Recognition Volume 172, Part A, April 2026, 112399

• Taehoon Kim,  Jaemin Na, Joong-won Hwang, Hyung Jin Chang, and Wonjun Hwang, Dual Prototype-driven Objectness Decoupling for Cross-Domain Object Detection in Urban Scene, Proceedings of the Asian Conference on Computer Vision (ACCV), 2024, pp. 1148-1165

• Dinh Phat Do, Taehoon Kim, Jaemin Na, Jiwon Kim, Keonho Lee, and Kyunghwang Cho, Wonjun Hwang, D3T: Distinctive Dual-Domain Teacher Zigzagging Across RGB-Thermal Gap for Domain-Adaptive Object Detection, In IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2024

• Seungbeom Woo, Geonwoo Baek, Taehoon Kim, Jaemin Na, Joong-won Hwang, and Wonjun Hwang, OurDB: Ouroboric Domain Bridging for Multi-Target Domain Adaptive Semantic Segmentation, In arXiv preprint arXiv:2403.11537, 2024

• Taehoon Kim, Jaemin Na,  Joong-Won Hwang, and Wonjun Hwang, Stay Focus On Object: Cross-Domain Detection using Domain-Invariant Object Representation, In IEEE International Conference on Image Processing (ICIP), 2024

• Hyeon Cho, Taehoon Kim, Hyung Jin Chang, and Wonjun Hwang, Self-Supervised Visual Learning by Variable Playback Speeds Prediction of a Video, IEEE Access, Vol. 9, pp. 79562-79571, May 2021 (IEEE Access), 2021

DOMESTIC CONFERENCES

• DIHN PHAT DO, Taehoon Kim, Wonjun Hwang, "MTR: Mix to Reduce Domain Gap Between RGB and Thermal for Domain-Adaptive Object Detection," 제 35회 인공지능신호처리학술대회, pp. 130-133, 2025 * Best Paper Award 

• Seungbeom Woo, Geonwoo Baek, Taehoon Kim, Jaemin Na, Joong-won Hwang, and Wonjun Hwang, "Recursive domain bridging for multi-target domain adaptive semantic segmentation", Image Processing and Image Understanding (IPIU), 2024. * Best Paper Award 

• Dinh Phat Do, Taehoon Kim, Jaemin Na, Jiwon Kim, Keonho Lee, and Kyunghwang Cho, Wonjun Hwang, "Distinctive Dual-Domain Teacher: Bridging the RGB-Thermal Gap in Domain Adaptive Object Detection", Image Processing and Image Understanding (IPIU), 2024. * Best Paper Award 

• Taehoon Kim, Jaemin Na, Wonjun Hwang, "Birds of a Feature: Cross-Domain Object Detection in Urban Scene Via Recycle Domain Invariant Knowledge", Image Processing and Image Understanding (IPIU), 2023. * Best Paper Award 

• Taehoon Kim, and Wonjun Hwang, “Improving Action Recognition task Using Co-training of Self-Supervised learning ” Summer Annual Conference of  The Institute of Electronics  and Information Engineers (IEIE), 2021 * Best Paper Award 

• Taehoon Kim and Wonjun Hwang, “Self-Supervised Spatiotemporal Learning For Video Using Variable Rotate Angle And Speed Prediction”, The Korean Institute of Broadcast and Media Engineers Summer Conference,(KIBME) 2020 

• Taehoon Kim, Wonjun Hwang and Hyeon Cho, “Semi-Supervised Classification Using Self-Supervised Technique”, Image Processing and Image Understanding (IPIU), 2020 

• Taehoon Kim and Wonjun Hwang, “Solving Jigsaw Puzzle Problem based on CNN”,  The Korean Institute of Communications  and Information Sciences Fall Conference (KICS), 2019 

---------------------------------------------------------------------------------------------------------------------------------------------------

Projects

• Developing the Knowledge Distillation technique for improving the performance of deep learning models, Hyundai Motor Company, Dec.2022 - Dec.2023

• Advanced Adaptive Learning Techniques for Networks Utilizing Complex environmental Data, Electronics and Telecommunications Research Institute (ETRI), Apr. 2023 – Dec. 2023

• Developing the Knowledge Distillation technique for improving the performance of deep learning models, Electronics and Telecommunications Research Institute (ETRI),  Apr. 2023 – Dec. 2023

• Research on unsupervised domain adaptation technology based on deep learning, Global Core Talent Training Support Project (IITP) , Sept. 2021 - Aug.2022

• Network technology research to adapt to various spatiotemporal complex situations , Electronics and Telecommunications Research Institute (ETRI), Apr. 2021 - Dec. 2022 

• Ajou DREAM Artificial Intelligence Innovative Talent Cultivation Project, Brain Korea (BK21): National Research Foundation of Korea, Sept. 2020 – Dec. 2023

• Development of deep learning-based situation recognition and judgment technology  for emergency steering control, South Korean Ministry of 

Trade Industry and Energy (MOTIE), May. 2018 - Present 

• Deep Learning-based Action Recognition using Self-supervised Learning, National Research Foundation of Korea (NRF), Mar. 2020. - Present

• Interaction force estimation for representation of physical senses based on image , Samsung Electronics, Sep. 2019 - Aug. 2020 

---------------------------------------------------------------------------------------------------------------------------------------------------