FastSurfer

FastSurfer processes structural human brain MRI scans on your own workstation or server, producing brain segmentations and cortical measurements compatible with FreeSurfer. It's for neuroimaging researchers working with individual scans or large cohorts who want shorter processing times within existing analysis workflows. The software is open source under Apache 2.0.

Starting with a T1-weighted scan, its PyTorch neural networks identify brain structures and calculate their volumes. The surface pipeline reconstructs the cortex and measures thickness across individual surface points and anatomical regions. Processing time varies with the enabled modules and available CPU or GPU hardware. The pipeline has been validated across public datasets for accuracy, reliability and sensitivity.

Specialized modules cover detailed cerebellum segmentation with CerebNet, hypothalamus and nearby structures with HypVINN, and corpus callosum thickness and shape measurements. HypVINN can also use a T2-weighted scan. FastSurfer-LIT uses a supplied lesion mask to support surface reconstruction in brains with tumors or cavities; its modified outputs need review before downstream analysis.

FastSurfer runs through Docker on workstations and servers, Apptainer on HPC clusters, or a native macOS package. It supports CPU-only processing and NVIDIA GPU acceleration, including a mixed mode that reduces GPU memory needs. Surface reconstruction requires a FreeSurfer license file; segmentation alone doesn't.