DeepBraTumIA is a deep learning based software tool for automated segmentation of glioblastoma multiform (GBM) multisequence MR images. As input the software requires the following MRI sequences: T1-weighted, T2-weighted, T1-contrast, FLAIR. The software enables automated skull stripping and voxel-wise segmentation of brain tumor and its main sub-compartments of necrotic, contrast-enhancing and edema. A batch mode is enabled for processing of many cases. DeepBraTumIA analyses hardware resources and adapts its memory and CPU/GPU requirements accordingly. Segmentations can be exported along with volumetric measurements of tumor compartments.
The tool can be used on pre- and post-operative cases, but better performance is known for pre-operative cases.