what is the best data augmentation for 3d brain tumor segmentation

• DOI: 10.48550/arxiv.2010.13372
• Publication Date: 2021-09-19
• Authors: Cirillo, Marco Do...; Abramian, David; Eklund, Anders
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Data augmentation; Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Computer Science - Computer Vision and Pattern Recognition; deep learning; Electrical Engineering and Systems Science - Image and Video Processing; 3D U-Net; artificial intelligence; Machine Learning (cs.LG); 3D brain tumor segmentation; 03 medical and health sciences; Medical Imaging; 0302 clinical medicine; Medicinsk bildvetenskap; FOS: Electrical engineering, electronic engineering, information engineering; Radiologi och bildbehandling; MRI; Radiology, Nuclear Medicine and Medical Imaging

Training segmentation networks requires large annotated datasets, which in medical imaging can be hard to obtain. Despite this fact, data augmentation has in our opinion not been fully explored for brain tumor segmentation. In this project we apply different types of data augmentation (flipping, rotation, scaling, brightness adjustment, elastic deformation) when training a standard 3D U-Net, and demonstrate that augmentation significantly improves the network's performance in many cases. Our conclusion is that brightness augmentation and elastic deformation work best, and that combinations of different augmentation techniques do not provide further improvement compared to only using one augmentation technique. Our code is available at https://github.com/mdciri/3D-augmentation-techniques.