Summary

Master's student in SKKU, advised by Jae-Pil Heo

Education

• Artificial Intelligence

  Present
  Sungkyunkwan University (SKKU)
• Electronic Electrical Engineering / Computer Science and Engineering (Double Major)

  2025.02
  Sungkyunkwan University (SKKU)

  GPA: 4.08/4.5

Work Experience

• Research Project Member

  2024-07 - Present
  IITP (Institute of Information & Communications Technology Planning and Evaluation)

  Research on improving intelligent command and control capabilities based on generative AI and real-time 3D digital twin construction

  • Semantic Segmentation Developer, 3D Reconstruction Developer
  • Drone Imagery based 3D Reconstruction (COLMAP & 3D Gaussian Splatting)
  • Remote-Sensing Image Semantic Segmentation
  • Open-Vocabulary Semantic Segmentation
• Research Project Member

  2023-10 - 2024-05
  Korean National Police Agency & Ministry of Science and ICT

  Detection of AI-based Fake Investigation and Tip Videos ($2.2M project)

  • Forgery Image Developer, Research Data Manager
  • Deepfake detection / Video forgery detection
  • Video inpainting / Video-to-video translation
• Co-Founder & Developer

  2020-05 - 2020-12
  Smart Maritime Logistics ICT Mentoring & Startup Audition

  Software development and startup management

  • Kalman filter based GNSS development
  • Android app development
  • Accounting and financial management
• Sergeant

  2021-05 - 2023-02
  Republic of Korea Air Force (ROKAF)

  Military service

  • Radar maintenance and monitoring duties
  • Routine inspections and troubleshooting for radar equipment

Skills

Publications

• Looking Beyond the Window: Global-Local Aligned CLIP for Training-free Open-Vocabulary Semantic Segmentation

  2026
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

  ByeongCheol Lee, Hyun Seok Seong, Sangeek Hyun, Gilhan Park, WonJun Moon, and Jae-Pil Heo

  View Publication
• PDF-GS: Progressive Distractor Filtering for Robust 3D Gaussian Splatting

  2026
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) [Findings]

  Kangmin Seo, MinKyu Lee, Tae-Young Kim, ByeongCheol Lee, JoonSeoung An, and Jae-Pil Heo

  View Publication

Languages

• Korean

  Native
• English

  Professional working proficiency

Interests

• Computer Vision

  Open-Vocabulary Semantic Segmentation, Domain-Generalized Semantic Segmentation, Dense Prediction, Zero-Shot Segmentation via VFM, Feature Upsampling, 3D Gaussian Splatting