Nikolai Filatov, Cels (CELS) and Peter the Great St. Petersburg Polytechnic University.
The author works at Cels on computer vision for chest CT and pelvic MRI, that is, on three-dimensional data, and is a PhD student at the Polytechnic University. The talk is about how to get value from diffusion generative models in medical computer vision with minimal effort, for real business tasks. Outline: the problem of data scarcity, what diffusion models are, how they are adapted for medicine, two practical cases (LoRA for pneumothorax and inpainting for rib fractures), and recommendations.
Continue reading — it's free
Google or a one-time email code. No password.
