Dipy

DIPY, short for Diffusion Imaging in Python, is an open source Python library for analysing diffusion MRI and other 3D and 4D medical imaging data. It brings together the full processing chain used in brain imaging research, from cleaning raw scans to reconstructing white matter pathways and running statistics on the results, and every algorithm is also available through command line interfaces. It is used by neuroimaging researchers, clinicians working in research settings and developers of imaging pipelines. Reconstruction covers single shell models such as DTI, CSA, Q Ball and CSD, and multi shell models such as DKI, GQI, SHORE, MAPMRI and MSMT CSD. Registration includes affine, diffeomorphic 2D and 3D, and streamline based methods. Tractography supports probabilistic, deterministic and particle filtering tracking. Denoising includes Patch2Self, a self supervised learning method, alongside non local means, Gibbs unringing and PCA based approaches. The library also handles brain extraction, SNR estimation and reslicing, provides statistical tools such as BUAN and AFQ, and includes interactive visualisation for tractograms and orientation distribution functions. Users can build their own algorithms on the same framework. DIPY is free to use. Professional support, consulting, custom development and training are offered through partner organisations listed on the project site, with pricing arranged directly with those partners. The project asks users to cite its 2014 paper in Frontiers in Neuroinformatics, and several of its methods were published in peer reviewed venues including NeurIPS, CVPR and NeuroImage.