Research Interests

• Computer Vision and deep learning.

• Continual learning, Open world object detection, Robotics, Foundations...

Education

• MS/Ph.D at Artificial Intelligence Graduate School, Ulsan National Institute of Science and Technology (UNIST), Ulsan, South Korea.

  • Advisor: Seungryul Baek

  • Mar. 2023 ~ Present

• B.S at Dept. of Information and Communications Engineering, Pukyong National University (PKNU), Busan, South Korea.

  • Advisor: Hoon Lee (Lab: ACNL)

  • Mar. 2016 ~ Feb. 2022.

Publications

• J. Kim, Y. Ku, J. Kim, J. Cha, S. Beak, "VLM-PL: Advanced Pseudo Labeling approach Class Incremental Object Detection with Vision-Language Model", in Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) workshop, Seattle, WA, 2024. (accepted) [PDF] 

• J. Kim, H. Cho, J. Kim, Y. Tiruneh, S. Beak, "SDDGR: Stable Diffusion-based Deep Generative Replay for Class Incremental Object Detection", in Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, 2024. (accepted) [PDF] (Highlight, 324/11532≈2.8% accept ratio) 

• J. Kim, S. Hong, C. Kim, J. Kim, Y. Tiruneh, J. On, J. Song, S. Choi, and S. Baek, "Class-Wise Buffer Management for Incremental Object Detection: An Effective Buffer Training Strategy," accepted for publication in Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Seoul, South Korea, 2024.[PDF]

• J. Kim, S. Hong, C. Kim, H. Cho, Yihalem, J. Kim, E. Kim, J. Song, S. Choi, S. Baek "An Effective Buffer Training Strategy for Class-Incremental Object Detection", non-archival, 2023 continual learning vision workshop(CLVision) in CVPRW, Jun. 2023. Subsequently, The improved version was accepted and will be published in ICASSP 2024.

• J. Kim, S. Hong, J. Kim, J. On, C. Kim, Yihalem, J. Song, S. Choi and S. Baek "Reducing data imbalancing for Object Detection", in Proc. 2023 Summer Conference of Institute of Electronics and Information Engineers (IEIE), Jun. 2023.

• H. Park, J. Kim, J. Kim and S. Baek, "Temporal MANO based Video Hand Pose and Shape Estimation", JKAIA, 2022.

• I. Lee, J. Kim, H. Lee, Y. Kim, D. Lee, H. Lee “Multi-Task Learning Approach to Autonomous Driving Techniques”, in Proc. 2021 Fall Conference of Institute of Electronics and Information Engineers (IEIE), Nov, 2021. (Best Paper Award)

• H. Kim, J. Kim, I. Lee, and H. Lee ”Deep Learning-Based Self-Driving Systems for Swarm Vehicle”, in Proc. 2021 Summer Conference of Korea Information and Communications Society (KICS), Jun. 2021. (Best Paper Award)

• J. Kim, H. Kim, I. Lee, H. Lee “AI Edge Server Design for Swarm Vehicular Network”, in Proc. 2021 ICT College Paper Contest of Institute of Embedded Engineering of Korea (IEMEK), May. 2021. (Best Paper Award)

• H. Kim, J. Kim, I. Lee, and H. Lee “AI-Enabled Low-Complexity Autonomous Driving Using Tensor RT”, in Proc. 2021 Winter Conference of Korea Information and Communications Society (KICS), Feb. 2021.

Submitted/Under-review

• Y. Jang*, J. Kim*, H. Kim, E. Lee, E. Kim, S. Baek, J. Yoo "B-RIGHT: Benchmark Re-evaluation for Integrity in Generalized Human-Object Interaction Testing", In under review. *co-authors 

Academic services

• Program Committee (Reviewer) at IEEE Transactions on Multimedia 2024

Experiences

• AI Ambassador, NVIDIA

  • • Feb. 2022 ~ Present

    • Mar. 2024 ~ Present. NVIDIA x 모두의 연구소 LAB 랩짱(Foundation Models LAB leader) 

    • official page

• AI Internship at Corp. KST, Busan, South Korea,

  • • Apr. 2022 ~ Jul. 2022

• AI researcher,  UVL Lab [homepage]

  • • Aug. 2022 ~ Feb. 2023

Projects

• "사람 중심 인공지능핵심원천기술개발"(IITP),  MIST(과기부),  Sep, 2023 - Present

• "Development of Object Detection Models for Impurity Detection in Fish Cake Products", Samjin-food, Sep. 2023 - Present

• "Expanding Data Coverage through Incremental Learning and Development of Integrated Domain Recognition Systems," LG Electronics, Aug. 2022 - Aug. 2023.

  • Recognized for achievements; the project led to the designation of our lab as a "Target Lab" in LG

• “Identification Systems of Dog Nose-Print Using Siamese Network, " Urobo AI(start-up) co-work, Sep. 2021 - Dec. 2021

• “AI-Assisted Safety Systems Using Industrial Internet-of-Things (IIoT), ” X-corps Festival, Apr.2021 - Dec.2021.

• “Autonomous Driving Techniques for Industrial Safety, ” Pukyong National University Han Contest, Jun. 2021 - Sep. 2021

• “Deep Learning Approaches for Swarm Autonomous Vehicle Systems, ”  Leaders in Industry-university Cooperation(LINC), Jan. 2021 - Mar. 2021.

Awards & Contests

•  Excellence Award, PKNU Study Club Performance, 2022, (Host: Pukyong Nation Univ.)

•  Excellence Award, LINK+ Competitive Contest, 2021, (Host: Pukyong Nation Univ.) 

•  Excellence Award, Han. Competitive Contest, 2021, (Host: Pukyong Nation Univ.) 

•  Best Undergraduate Paper Award, 2021, (Host: IEMEK, The Institute of Embedded Engineering of Korea) 

•  Best Undergraduate Paper Award, 2021, (Host: KICS, The Korea’s information and communications technology) 

•  Best Undergraduate Paper Award, 2021, (Host: IEIE, The Institute of Electronics and Information Engineers) 

•  Excellence Award, Presentation Engineering in literature, 2021, (Host: Pusan National Univ.) 

•  1st Prize, PKNU Capstone Design, 2021, (Host: Pukyong Nation Univ) 

• 2nd Prize (KIBO), Busan AI Idea Contest Exhibition, 2020, (Host: NILE, Busan) 

Affiliations

• Notion - Official offered by notion team (Jan. 2023 ~ present)

TA & Workshops

• NVIDIA Jetson AI Ambassador

  • Workshop presentation at PKNU, May. 2022 [LINK]

  • Workshop presentation at Corp. KST, May. 2022

• Novatus

  • 5th AI Novatus Academia (Computer Vision part) at UNIST AI Innovation Park, Masan, South Korea, 28 Apr. 2023. 

  • 6th AI Novatus Academia (Computer Vision part) at UNIST AI Innovation Park, Ulsan, South Korea, 25 Aug. 2023. 

• Continual Learning Vision Workshop(CLVision), CVPR 2023,

  • Poster presentation

  • An Effective Buffer Training Strategy for Class-Incremental Object Detection

 Contents

• Youtube presentation

  • Anydoor / GLIGEN+BLIP / LayoutDM / Classifier-free guidance / I-blurry /