Byeonghun Kim

I am a Master student at Department of Future Convergence Technology in Soonchunhyang Univ. and advised by Byeongjoon Noh.

My research interests lie in Computer Vision, Federated Learning, and Action Recognition.

Recently, I have been interested in applying large language model (LLM) and multi-modal large language models (MLLM) to autonomous driving and intelligent transportation systems.

My research primarily focuses on applying deep learning and machine learning to improve traffic safety and predict pedestrian risk through computer vision technology.

Email: byeonghun@sch.ac.kr

Phone: +82)10-5015-5129

News

Jun 05, 2025	Elsevier SCI(E)급 저널 리뷰어 위촉 – JRESS(IF: 9.4), ESWA(IF: 7.5)
Mar 18, 2025	Published in Applied Intelligence (SCI)
Mar 04, 2025	Oral presentation at AAAI 2025 FLUID Workshop
Aug 29, 2024	한국연구재단 석사과정생연구장려금지원사업 선정

Latest posts

Apr 11, 2025	[Paper Review] A Review of DeepSeek Models' Key Innovative Techniques
Nov 14, 2024	[Paper Review] Can Language Beat Numerical Regression? Language-Based Multimodal Trajectory Prediction
Sep 23, 2024	[Paper Review] EAGLE: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders

Latest publications

International Jour.

Federated learning-based road surveillance system in distributed CCTV environment: Pedestrian fall recognition using spatio-temporal attention networks

Byeonghun Kim, Jaegyun Im, and Byeongjoon Noh

Applied Intelligence, 2025

DOI Bib

@article{kim2025federated,
  title = {Federated learning-based road surveillance system in distributed CCTV environment: Pedestrian fall recognition using spatio-temporal attention networks},
  author = {Kim, Byeonghun and Im, Jaegyun and Noh, Byeongjoon},
  journal = {Applied Intelligence},
  volume = {55},
  number = {6},
  pages = {1--16},
  year = {2025},
  doi = {10.1007/s10489-025-06451-z},
  publisher = {Springer},
}

International Conf.

FLAMe: Federated Learning with Attention Mechanism using Spatio-Temporal Keypoint Transformers for Pedestrian Fall Detection in Smart Cities

Byeonghun Kim, and Byeongjoon Noh

AAAI 2025 FLUID Workshop, 2025

DOI Bib

@article{kim2024flame,
  title = {FLAMe: Federated Learning with Attention Mechanism using Spatio-Temporal Keypoint Transformers for Pedestrian Fall Detection in Smart Cities},
  author = {Kim, Byeonghun and Noh, Byeongjoon},
  journal = {AAAI 2025 FLUID Workshop},
  year = {2025},
  doi = {10.48550/arXiv.2412.14768},
  url = {https://arxiv.org/abs/2412.14768},
}

International Jour.

Method on Efficient Operation of Multiple Models for Vision-Based In-Flight Risky Behavior Recognition in UAM Safety and Security

Byeonghun Kim, Byeongjoon Noh, and Kyowon Song

Journal of Advanced Transportation, 2024

DOI Bib

@article{kim2024method,
  title = {Method on Efficient Operation of Multiple Models for Vision-Based In-Flight Risky Behavior Recognition in UAM Safety and Security},
  author = {Kim, Byeonghun and Noh, Byeongjoon and Song, Kyowon},
  journal = {Journal of Advanced Transportation},
  volume = {2024},
  number = {1},
  pages = {7113084},
  year = {2024},
  publisher = {Wiley Online Library},
  doi = {10.1155/2024/7113084},
  url = {https://onlinelibrary.wiley.com/doi/full/10.1155/2024/7113084},
}