Research Intern at EMI Lab, Korea University, and a B.S. candidate in Artificial Intelligence at Gachon University.
I am interested in long video understanding and efficient visual perception. My research background includes efficient object detection and real-time high-resolution perception systems.
Long Video Understanding · Computer Vision · Efficient Object Detection · Real-Time High-Resolution Perception · Multimodal Perception
My hands-on experience includes:
- training and evaluating object detectors with PyTorch, YOLO, and DETR-style architectures
- optimizing high-resolution / small-object detection pipelines with OpenCV, TensorRT-FP16, and GPU workflows
- building Android perception prototypes with Kotlin, CameraX, TFLite/LiteRT, TTS, and haptic feedback
- running experiments on Linux workstations with Git, SSH, Jupyter, and CUDA-based tooling
- SQ-DETR: DETR-style object detection research code related to layer-wise query selection and efficient detection.
- ASAP-Bird-Detection: asynchronous slicing pipeline for real-time high-resolution small-bird detection.
- SmartMedia-YOLO26-Small-Object-Experiments: small-object detection ablation experiments using YOLO26-family models.
- VIA-Safe-Crosswalk: Android assistive crosswalk guidance app using CameraX, on-device traffic-light detection, TTS, and haptic feedback.
- Layer-Wise Query Selection to Eliminate Redundant Queries in DETR
Applied Sciences, 2025-07-09. DOI: 10.3390/app15147686
Contributors: Seok-Jin Hong, Chan-Young Choi, Sang-Woong Lee · Co-first author. - Asynchronous Slicing Accelerated Pipeline for Real-time High-Resolution Small Bird Flock Monitoring — under peer review. Code: ASAP-Bird-Detection
- Email: kinetic27@gachon.ac.kr / aheui@kakao.com
