Kitware released version 4.11 of the National Library of Medicine Insight Segmentation and Registration Toolkit (ITK) to support scientific discovery through medical image analysis. Version 4.11 added remote modules to increase support for the capabilities of the toolkit. This release also bolsters code coverage, and provides members of the medical community with an open-source, cross-platform solution for analyzing images such as those acquired from computed tomography, magnetic resonance imaging, and ultrasound procedures.
Remote modules served as a noteworthy area of focus for ITK 4.11 development. With new modules, ITK perturbs mesh vertices with additive Gaussian noise; accurately benchmarks computational performance; performs principal components analysis of scalar, vector, and mesh vertex data; interpolates multi-label images; and iterates over multiple images simultaneously. Several other improvements made to ITK pertain to filtering. With the addition of UnsharpMaskImageFilter, for example, enhancing edges is as simple as ever.
ITK also improved input/output with Neuroimaging Informatics Technology Initiative (NIfTI) support, and it enhanced wrapping. ITK now wraps more classes and uses a new CMake function to test Python scripts in modules. Furthermore, ITK learned to only load individual Python modules when required, which decreases load time. In addition, ITK updated third-party libraries such as the vision numerics library (VNL) and Digital Imaging and Communications in Medicine Toolkit (DCMTK); improved code coverage to over 85 percent; fixed bugs; and increased documentation with modifications to The ITK Software Guide, Doxygen text, Wiki page, and Sphinx examples.
The release featured the work of over 30 community members, including nine first-time contributors. For more information on ITK 4.11 and to download the toolkit, please visit https://itk.org. Beyond documentation, Kitware offers consulting and support services for ITK. To learn how to leverage these services, please contact firstname.lastname@example.org.