Voxel is a broad framework for hardware-accelerated volumetric medical imaging, centered on image and mesh data structures and their relationship to a world coordinate system. It is built on PyTorch, so data manipulations run on the GPU and support automatic differentiation, making it well-suited for rapid algorithm development, learning-based pipelines, or simply fast inspection of imaging data in Python.
The library provides data structures and torch-like operations spanning 3D image data and acquisition geometry, resampling, filtering, and morphology, segmentation and label maps, affine and nonlinear transformations and registration, triangular mesh processing, and broad medical imaging I/O. Install the latest release with pip.
pip install voxel
Guides and the full API reference live in the voxel documentation. Pull requests and bug reports are welcome.
If voxel is useful in your work, please cite the publication it was originally developed for.
VoxelPrompt: A Vision Agent for End-to-End Medical Image Analysis
Andrew Hoopes, Neel Dey, Victor Ion Butoi, John V. Guttag, Adrian V. Dalca