Hongje Seong

hongjeseong.jpg

📧 hongje.seong [at] samsung.com

📍 130, Samsung-ro, Yeongtong-gu, Suwon-si, Gyeonggi-do, 16678, Korea

CV | Google Scholar | Github

I am a researcher at Samsung Advanced Institute of Technology (SAIT).

I received my Ph.D. and B.S. degrees in Electrical and Electronic Engineering from Yonsei University in 2023 and 2018, respectively, advised by Professor Euntai Kim. I was a research intern at Adobe Research (San Jose, CA) in 2021. During my internship, I worked at the Creative Intelligence Lab with Joon-Young Lee, Seoung Wug Oh, and Brian Price as my mentors.

I am broadly interested in computer vision. Specifically, my research interests include

  • Autonomous Driving
  • Image / Video Segmentation
  • Image / Video Matting
  • Image / Video Recognition
  • Domain Adaptation / Generalization
  • Image Retrieval / Place Recognition
but not limited to.

selected publications

  1. Revisiting Self-Similarity: Structural Embedding for Image Retrieval
    Seongwon Lee, Suhyeon Lee,  Hongje Seong, and Euntai Kim
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun, 2023
    (Acceptance: 2360/9155 ≈ 25.8%)
    Page
  2. Fallen Person Detection for Autonomous Driving
    Suhyeon Lee, Sangyong Lee,  Hongje Seong, Junhyuk Hyun, and Euntai Kim
    Expert Systems With Applications (ESWA), Mar, 2023
    (IF: 8.665 in JCR2021)
    Page
  3. Video Object Segmentation using Kernelized Memory Network with Multiple Kernels
    Hongje Seong, Junhyuk Hyun, and Euntai Kim
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), Feb, 2023
    (IF: 24.314 in JCR2021)
    Page
  4. SHUNIT: Style Harmonization for Unpaired Image-to-Image Translation
    Seokbeom Song, Suhyeon Lee,  Hongje Seong, Kyoungwon Min, and Euntai Kim
    AAAI Conference on Artificial Intelligence (AAAI), Feb, 2023
    (Acceptance: 1721/8777 ≈ 19.6%)
    arXiv
  5. Domain Adaptive Video Semantic Segmentation via Cross-Domain Moving Object Mixing
    Kyusik Cho, Suhyeon Lee,  Hongje Seong, and Euntai Kim
    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Jan, 2023
    (Acceptance: 191/623 ≈ 30.7% in Round 2)
    arXiv Page Code
  6. One-Trimap Video Matting
    Hongje Seong, Seoung Wug Oh, Brian Price, Euntai Kim, and Joon-Young Lee
    European Conference on Computer Vision (ECCV), Oct, 2022
    (Acceptance: 1645/5804 ≈ 28.3%)
      Winner at Qualcomm Innovation Fellowship 2022  
    arXiv Page Code Video
  7. Adjacent Feature Propagation Network (AFPNet) for Real-Time Semantic Segmentation
    Junhyuk Hyun,  Hongje Seong, Sangki Kim, and Euntai Kim
    IEEE Transactions on Systems, Man, and Cybernetics: Systems (TSMC), Sep, 2022
    (IF: 11.471 in JCR2021)
    Page
  8. Indoor Place Category Recognition for a Cleaning Robot by Fusing a Probabilistic Approach and Deep Learning
    Soowook Choe*,  Hongje Seong*, and Euntai Kim
    IEEE Transactions on Cybernetics (TCYB), Aug, 2022
    (IF: 19.118 in JCR2021, * indicates equal contribution)
    Page
  9. Correlation Verification for Image Retrieval
    Seongwon Lee,  Hongje Seong, Suhyeon Lee, and Euntai Kim
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun, 2022
    (Acceptance: 342/8161 ≈ 4.2%)
      Oral Presentation  
      Winner at Qualcomm Innovation Fellowship 2022  
    arXiv Page Code
  10. WildNet: Learning Domain Generalized Semantic Segmentation from the Wild
    Suhyeon Lee,  Hongje Seong, Seongwon Lee, and Euntai Kim
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun, 2022
    (Acceptance: 2067/8161 ≈ 25.3%)
      Finalist at Qualcomm Innovation Fellowship 2022  
    arXiv Page Code
  11. Iteratively Selecting an Easy Reference Frame Makes Unsupervised Video Object Segmentation Easier
    Youngjo Lee,  Hongje Seong, and Euntai Kim
    AAAI Conference on Artificial Intelligence (AAAI), Feb, 2022
    (Acceptance: 1349/9020 ≈ 15.0%)
    arXiv Page
  12. Graph-Based Point Tracker for 3D Object Tracking in Point Clouds
    Minseong Park,  Hongje Seong, Wonje Jang, and Euntai Kim
    AAAI Conference on Artificial Intelligence (AAAI), Feb, 2022
    (Acceptance: 1349/9020 ≈ 15.0%)
    Page
  13. Universal Pooling - A New Pooling Method for Convolutional Neural Networks
    Junhyuk Hyun,  Hongje Seong, and Euntai Kim
    Expert Systems With Applications (ESWA), Oct, 2021
    (IF: 6.954 in JCR2020)
    Page
  14. Hierarchical Memory Matching Network for Video Object Segmentation
    Hongje Seong, Seoung Wug Oh, Joon-Young Lee, Seongwon Lee, Suhyeon Lee, and Euntai Kim
    IEEE/CVF International Conference on Computer Vision (ICCV), Oct, 2021
    (Acceptance: 1617/6236 ≈ 25.9%)
    arXiv Page Code Video
  15. Unsupervised Domain Adaptation for Semantic Segmentation by Content Transfer
    Suhyeon Lee, Junhyuk Hyun,  Hongje Seong, and Euntai Kim
    AAAI Conference on Artificial Intelligence (AAAI), Feb, 2021
    (Acceptance: 1692/7911 ≈ 21.4%)
    arXiv Page
  16. Kernelized Memory Network for Video Object Segmentation
    Hongje Seong, Junhyuk Hyun, and Euntai Kim
    European Conference on Computer Vision (ECCV), Aug, 2020
    (Acceptance: 1361/5025 ≈ 27.1%)
      3rd Place Award in DAVIS20 Challenge (CVPRW-DAVIS 2020)  
    arXiv Page Video