I'm a Master's student in Software Engineering at Northwestern Polytechnical University (NPU), China (CSC Scholar), working on medical image analysis — specifically semi-supervised tumor segmentation and foundation models for medical imaging.

Before my Master's, I completed my Bachelor's in Bangladesh and spent ~2 years as a software engineer, so I care about research that is both rigorous and reproducible.

🔬 Research focus: I work on label-efficient medical image segmentation — reducing the heavy annotation cost of 3D medical imaging by combining semi-supervised learning with vision foundation models (SAM/MedSAM). My current work uses confidence-gated pseudo-labeling to make foundation-model refinement reliable enough for tumor segmentation under very few labels.

📄 Recent work: Semi-Supervised Tumor Segmentation via MedSAM2 Confidence-Gated Pseudo-Label Refinement — under review, IEEE Transactions on Biomedical Engineering (TBME), 2026. → Code

🎯 Currently seeking a fully funded PhD position (2027) in medical image computing.

• 📧 Email: shanta.0713islam@gmail.com
• 💼 LinkedIn: https://www.linkedin.com/in/shantaislambangladesh/

Python · PyTorch · nnU-Net · 3D Medical Image Segmentation · Semi-Supervised Learning · Foundation Models (SAM / MedSAM2) · Pseudo-Labeling · Uncertainty / Confidence Estimation · Deep Learning