GPU Acceleration, Rapid Releases, and Biomedical Examples for scikit-image
Gregory Lee (Quansight)
To maintain the popular scikit-image Python library for microscopy and medical imaging data and bring significant improvements via development of a backend system enabling multi-threading and GPU acceleration, an improved release process for more rapid cycles, and expanded biomedical examples.
scikit-image is the open source image processing toolkit of scientific Python. It proposes a collection of algorithms which address the various image processing tasks encountered in science (denoising, segmentation, feature extraction). scikit-image is application-agnostic, and its algorithms accept both two-dimensional and three-dimensional (sometimes n-d) images for compatibility with various scientific image modalities, including microscopy, tomography, and MRI. However, it is a core dependence of many application-specific image processing packages such as CellProfiler or hyperspy. scikit-image targets a wide community of students, engineers, and scientists, many of them self-taught about image processing.