Avatar

Junho Koh

Ph.D Candidate

Hanyang University


CV | Google Scholar | LinkedIn | Github


I am a Ph.D candidate at Hanyang University (HYU), advised by Professor Jun Won Choi.

My research interest lies in perception for autonomous vehicles and mobile robots. Particularly, I focus on 2D or 3D object detection, sensor fusion-based perception, object tracking, video object detection, and other related areas.

I am a recipient of Hyundai Motor Group Research Scholarship .

Contact

  • jhkoh [at] spa.hanyang.ac.kr

  • Room #408-1, Engineering Center Main Building, Hanyang University, Seoul, Republic of Korea, 04763

Education

  • PhD in Electrical Engineering, 2024

    Hanyang University, Seoul

  • MS in Electrical Engineering, 2020

    Hanyang University, Seoul

  • BS in Electrical Engineering, 2018

    Hanyang University, Seoul

Publications

    * indicates equal contributions

  • Fine-grained pillar feature encoding via spatio-temporal virtual grid for 3D object detection

    Konyul Park*, Yecheol Kim*, Junho Koh, Jun Won Choi

    IEEE International Conference on Robotics and Automation (ICRA), 2024.

    [ Coming soon... ]

  • Unified contrastive fusion transformer for multimodal human action recognition

    Kyoung Ok Yang, Junho Koh, Jun Won Choi

    Preprint

    [ Paper ]

  • MGTANet: Encoding sequential LiDAR points using long short-term motion-guided temporal attention for 3D object detection

    Junho Koh*, Junhyung Lee*, Youngwoo Lee, Jaekyum Kim, Jun Won Choi

    Association for the Advancement of Artificial Intelligence (AAAI), 2023 (Acceptance: 1,721/8,777 < 20%) [Oral]

    Ranked 3rd place among LiDAR methods on nuScenes object detection benchmark as of August 2022

    [ Paper ]

  • D-Align: Dual query co-attention network for 3D object detection based on multi-frame point cloud sequence

    Junhyung Lee, Junho Koh, Youngwoo Lee, Jun Won Choi

    IEEE International Conference on Robotics and Automation (ICRA), 2023.

    [ Paper ]

  • Joint 3D object detection and tracking using spatio-temporal representation of camera image and LiDAR point clouds

    Junho Koh*, Jaekyum Kim*, Jinhyuk Yoo, Yecheol Kim, Dongsuk Kum, Jun Won Choi

    Association for the Advancement of Artificial Intelligence (AAAI), 2022 (Acceptance: 1,349/9,251 < 15%)

    [ Paper ]

  • Joint representation of temporal image sequences and object motion for video object detection

    Junho Koh*, Jaekyum Kim*, Younji Shin, Byeongwon Lee, Seungji Yang, Jun Won Choi

    IEEE International Conference on Robotics and Automation (ICRA), 2021 [Oral]

    [ Paper ]

  • Video object detection using object's motion context and spatio-temporal feature aggregation

    Jaekyum Kim*, Junho Koh*, Jun Won Choi

    IEEE International Conference on Pattern Recognition (ICPR), 2020

    [ Paper ]

  • Enhanced object detection in bird's eye view using 3D global context inferred from Lidar point data

    Yecheol Kim, Jaekyum Kim, Junho Koh, Jun Won Choi

    IEEE Intelligent Vehicles Symposium (IV), 2019

    [ Paper ]

  • Robust deep multi-modal learning based on gated information fusion network

    Jaekyum Kim, Junho Koh, Yecheol Kim, Youngbae Hwang, Jun Won Choi

    Asian Conference on Computer Vision (ACCV), 2018

    [ Paper ]

  • Robust camera lidar sensor fusion via deep gated fusion network

    Jaekyum Kim, Jaehyung Choi, Yecheol Kim, Junho Koh, Chung Choo Chung, and Jun Won Choi

    IEEE Intelligent Vehicles Symposium (IV), 2018

    Among 5% selected as single track oral presentation

    [ Paper ]

Activities

Reviewer

  • IEEE Transactions on Vehicular Technology
  • IEEE Transactions on Intelligent Transportation Systems
  • Neurocomputing

Talks

  • AAAI2023 Review Seminar, EIRIC

    Date: May, 2023
    Topic: LiDAR-based 3D Object Detection
    [Page | Slide | Video]

Honors

  • Research Scholarship Student, Hyundai Motor Group ($ 1,700 per month)
    Jun, 2021-Dec, 2023
  • ICT Talent Development Scholarship, Huawei ($ 2,000)
    Oct, 2022