Computer Vision Researcher · Medical Image Analysis · Hanoi University of Science and Technology
I am a Computer Vision researcher at Hanoi University of Science and Technology (HUST), working primarily on medical image analysis.
My research interests include cardiac MRI segmentation, physics-guided deep learning, efficient vision architectures, and 2D-to-3D medical image reconstruction. I am particularly interested in building models that remain reliable under imperfect data and practical deployment constraints.
Research focus: Medical Image Segmentation · Cardiac MRI · Physics-Informed Learning · 3D Reconstruction · Efficient Vision Models
Integrating physics guidance into deep learning for cine cardiovascular segmentation
My Anh Tran-Nguyen, Quoc Khanh Luong, Minh Bao Kha, Tuan Linh Dang.
Computerized Medical Imaging and Graphics, 134, 102798 (2026).
Presents PGE-UNet, a lightweight cine-CMR segmentation framework combining a noise-aware encoder, physics-regularized decoder, and simulated RF transmit-field guidance.
Read the paper · Source code · Model weights
| Project | Focus |
|---|---|
| PGE-UNet | Published physics-guided, parameter-efficient segmentation for cine cardiac MRI. |
| Self-Audit / SpecUMamba | Ongoing research into image-only cardiac segmentation, conservative pseudo-labeling, and student refinement without ground-truth supervision during training. |
MONAI · NumPy · Scientific Computing · Reproducible ML
3D contribution calendar · automatically refreshed by GitHub Actions
Recent GitHub activity
Last 31 days · automatically updated daily by GitHub Actions
GitHub · LinkedIn · Facebook · Instagram
Interested in medical imaging, robust vision models, or research collaboration? Feel free to reach out on LinkedIn.