Hello! I’m **Junho Lee**, an undergraduate student in the **Department of Computer & Artificial Intelligence** at **Jeonbuk National University (JBNU)**.
I am currently conducting research at **MacsLAB (Medical AI Laboratory)**, focusing on **medical image analysis** and **deep learning model development**.
My main interests include **improving AI model accuracy**, **automating medical data preprocessing**, and **developing web-based medical image viewers**.
I primarily work with PyTorch for modeling and FastAPI + Vue3 (TypeScript) for full-stack development.
To make AI models deployable in real clinical environments,
I design and implement the entire pipeline—from backend to frontend to AI model integration—myself.
Balancing research and development, I aim not only to enhance performance but also to design practically usable systems.
In particular, I focus on improving DICOM viewer UX/UI, AI result visualization, and interactive measurement tools,
developing interfaces that can be applied directly in medical practice.
🔬 Main Research & Projects
- DICOM Viewer Development: Displays medical images (CT, X-ray, etc.) with distance, angle, and area measurement tools and window-level adjustment functions
- AFF (Atypical Femoral Fracture) Detection Model: Built PyTorch-based CNN model and Grad-CAM visualization pipeline
- Medical Image Analysis Web Platform: Integrated FastAPI backend and Vue3 frontend, deployed on the cloud
- MNIST Classification Model: Trained MLP model for handwritten digit (0–9) recognition and developed a visualization dashboard
- Automatic DICOM Metadata Parsing & Pixel Scaling: Implemented precise measurements using actual physical units (mm/px)
🧭 Future Goals
- Enhance explainability of AI models through visualization research
- Commercialize and optimize web-based diagnostic assistance systems for medical imaging
- Build a general-purpose medical imaging platform applicable to various medical datasets
🔗 Quick Links