☕ 마신 커피 만큼 코드를 뽑아냅니다. 👨💻
B.S. candidate in Artificial Intelligence at Gachon University.
Undergraduate researcher interested in computer vision and real-time perception systems.
ORCID: 0009-0003-3938-1832
Computer Vision · Object Detection · Efficient / Lightweight Detection · Real-Time High-Resolution Vision Systems · Small Object Detection · Robust & Multimodal Perception
I mostly build and experiment with vision models and practical perception systems:
- 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 / experiment repository related to layer-wise query selection and efficient detection.
- 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.
GitHub stats & extras
- Discord DM: factorio.com
- Email: kinetic27@gachon.ac.kr / aheui@kakao.com
