Publications

You can also find my papers on Google Scholar.

Selected Publications

• Leveraging Temporal Contextualization for Video Action Recognition
  Minji Kim, Dongyoon Han, Taekyung Kim, Bohyung Han
  ECCV 2024 Temporal Contextualization (TC) enhances CLIP’s video understanding capability by infusing temporal information through global interactions within the encoding process.

• Towards Sequence-Level Training for Visual Tracking
  Minji Kim*, Seungkwan Lee*, Jungseul Ok, Bohyung Han, Minsu Cho (*Equal Contribution)
  ECCV 2022
  Sequence-Level Training (SLT) resolves the training-testing inconsistency problem of recent object trackers by directly optimizing a test-time metric under a reinforcement learning framework.

• Online Hybrid Lightweight Representations Learning: Its Application to Visual Tracking
  Ilchae Jung, Minji Kim, Eunhyeok Park, Bohyung Han
  IJCAI 2022
  A hybrid representation learning from low-bit quantized and channel-pruned networks improves online tracker updates while reducing computational costs.

Earlier Projects (~2021)

Below works were mostly done during my undergraduate research internship and the early years of graduate studies, with a focus on deformable object tracking and tracker model ensemble.

• Learning Gaussian Models for Orientation-Aware Visual Tracking
  Minji Kim, Bohyung Han
  US Patent (App. 18/370,531)

• Top-down Thermal Tracking Based on Rotatable Elliptical Motion Model for Intelligent Livestock Breeding
  Minji Kim, Wonjun Kim
  Multimedia Systems, Springer, 2020

• Ensemble Modeling with Convolutional Neural Networks for Application in Visual Object Tracking
  Minji Kim, Ilchae Jung, Bohyung Han
  Journal of Korean Institute of Information Scientists and Engineers (KIISE), 2020