nnU-Net: Automated Design of Deep Learning Methods for Biomedical Image Segmentation

Video

Abstract: Biomedical imaging is a driver of scientific discovery and core component of medical care, currently stimulated by the field of deep learning. While semantic segmentation algorithms enable 3D image analysis and quantification in many applications, the design of respective specialised solutions is non-trivial and highly dependent on dataset properties and hardware conditions. We propose nnU-Net, a deep learning framework that condenses the current domain knowledge and autonomously takes the key decisions required to transfer a basic architecture to different datasets and segmentation tasks. Without manual tuning, nnU-Net surpasses most specialised deep learning pipelines in 19 public international competitions and sets a new state of the art in the majority of the 49 tasks. The results demonstrate a vast hidden potential in the systematic adaptation of deep learning methods to different datasets. We make nnU-Net publicly available as an open-source tool that can effectively be used out-of-the-box, rendering state of the art segmentation accessible to non-experts and catalyzing scientific progress as a framework for automated method design.

Paul F. Jaeger is a PostDoc in Medical Image Analysis at the German Cancer Research Center (DKFZ). He studied Physics in Karlsruhe, Stockholm, and Melbourne. During his PhD at KIT and DKFZ he spent 6-months as a Research Intern at Facebook AI Research in Montreal and initiated heidelberg.ai, a platform connecting over 1600 AI-enthusiasts in the area. His research interests lie in the intersection of Object Detection and Semantic Segmentation with a focus on making Deep Learning models succeed in real-life clinical applications. Paul is leading the Research Working Group at MONAI, an international initiative committed to advancing the field of Medical Image Analysis by implementing standards for reproducibility, transparency, and method evaluation.