Invited Seminar: Applications of Advanced Deep Learning for Prostate Segmentation

This seminar presents applications of advanced deep learning models for accurate prostate segmentation using T2-weighted MRI datasets of prostate cancer patients. Techniques include U-Net, ERF-Nets, GANs, Transformer-based models, and Diffusion models applied to medical imaging and radiomics.

The seminar demonstrates how these models improve segmentation accuracy, enhance tissue boundary detection, and enable reliable extraction of radiomic features for downstream clinical applications.

Key topics include:

• Preprocessing and augmentation of T2-weighted MRI prostate datasets
• Training and evaluation of GANs, Transformers, Diffusion, and CNN-based segmentation models
• Integration of segmentation results into radiomic pipelines for prostate cancer characterization
• Challenges and opportunities for generative AI in clinical imaging