3.0

credits

Average Course Rating

This project-based research course explores generative models in the context of computer vision and biomedical imaging, focusing on both theoretical foundations and practical applications. Topics include variational autoencoders (VAEs), generative adversarial networks (GANs), diffusion models, and multimodal generation (e.g., text-to-image, image-to-text, video generation). Students will study state-of-the-art research, implement generative models, and apply them to real-world visual data. The course emphasizes understanding underlying probabilistic principles, architectures, and training challenges. A semester-long project will allow students to apply state-of-the-art models to a real-world vision problem, with special attention to biomedical imaging applications such as medical image segmentation, synthesis, denoising, restoration, and anomaly detection for diagnostics.