nnDetection: A Self-configuring Method for Medical Object Detection

Simultaneous localisation and categorization of objects in medical images, also referred to as object detection, is high clinical relevance because diagnostic decisions often depend on rating rather than e.g. pixels. For this task, the cumbersome iterative process method configuration constitutes a major research bottleneck. Recently, nnU-Net has tackled challenge for task image segmentation with great success. Following nnU-Net's agenda, work we systematize automate detection. The resulting self-configuring method, nnDetection, adapts itself without any manual intervention arbitrary detection problems while achieving results en par or superior state-of-the-art. We demonstrate effectiveness nnDetection two public benchmarks, ADAM LUNA16, propose 11 further tasks data sets comprehensive evaluation. Code at https://github.com/MIC-DKFZ/nnDetection .