Huge thanks to Jin's co-supervisor Dan Marcus, his PhD committee, collaborators and the Imaging Science PhD Program. We’re immensely proud of Jin’s hard work, rigor, and perseverance. We are excited to see what he will accomplish next . 👏
#PhDDefense #DeepLearning #ProudAdvisor

📄 ASFDA (#MICCAI2025):
Investigates Active Source-Free Domain Adaptation for medical segmentation.
Selects the most informative target samples via Image Sensitivity & Organ Heterogeneity queries, boosting performance without access to source data.
t.co/qu0tHpSiQI

📄D-Net (Biomed Signal Process Control 2026):
A novel Dynamic Large-Kernel Transformer integrating adaptive feature fusion for 3D organ and tumor segmentation. Superior segmentation performance across several tasks, with lower computational complexity.
doi.org/10.1016/j.bspc

📄 DMC-Net (Biomed Signal Process Control 2025):
Introduced Dynamic Multi-Scale and Multi-Resolution Convolutions to better capture organ variability.
Achieved state-of-the-art pancreas segmentation while keeping computation lightweight and efficient.
doi.org/10.1016/j.bs...

His thesis advanced the field of deep learning for medical image segmentation, developing high-performance, data-efficient methods for organ and lesion segmentation and for adapting models across imaging domains and modalities.
📚 Selected publications from his thesis work👇