Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
This lecture from the MedAI Group Exchange Sessions features Zheyuan Zhang discussing robust medical image analysis challenges and solutions. Learn about the problems of data distribution differences across medical centers and how to develop models that can generalize to previously unseen domains. Explore Zhang's work on collecting a large multi-center abdominal MRI dataset, developing PanSegNet (a new pancreas segmentation method combining nnUNet with Transformer networks), implementing domain generalization through adversarial intensity attacking, creating controllable diffusion models for medical image synthesis, and constructing semi-supervised algorithms that leverage unlabeled data for large-scale medical image segmentation. Zhang, who earned his bachelor's from Tsinghua University and pursued his Ph.D. at Northwestern University, shares insights on how recent large model advancements can help create robust solutions for real-world clinical challenges in medical imaging.
