ITK-SNAP

ITK-SNAP is a free software application for semi-automatic and manual segmentation of structures in three-dimensional and four-dimensional biomedical images. Originally developed at the University of North Carolina, the software combines active contour methods (snakes) with manual delineation tools to facilitate quantitative analysis in clinical and preclinical research. The project is currently maintained by the Penn Image Computing and Science Laboratory (PICSL) at the University of Pennsylvania.

★★★★★

5.0(1 ratings)

• Synchronized multiplanar navigation. The linked cursor enables simultaneous movement through the axial, coronal, and sagittal planes. The three orthogonal views update in real time, facilitating precise localization of anatomical structures in three-dimensional space.
• Manual segmentation in three planes. The user can delineate regions of interest by painting or tracing contours on any of the orthogonal planes. Edits made on one plane are immediately reflected on the other two, allowing exhaustive control of the segmentation.
• Semi-automatic region growing segmentation. The active contour (snake) algorithm evolves from an initial seed to fit the boundaries of the target structure. Parameters such as stiffness, curvature, and intensity can be adjusted for each specific task.
• Qt6-based graphical user interface. The modern interface includes dockable panels, customizable keyboard shortcuts, and support for high-resolution monitors. All controls are organized to minimize cognitive load during the workflow.
• Support for multiple image formats. ITK-SNAP can read and write formats such as NIfTI (.nii, .nii.gz), DICOM (complete series or single files), Analyze, MetaImage, NRRD, and TIFF. DICOM directory import automatically organizes series by modality and orientation.
• Concurrent visualization and segmentation of multiple images. Multiple images can be loaded simultaneously (e.g., T1, T2, FDG-PET) and kept aligned through the same coordinate system. Segmentations performed on one image can be overlaid onto the others.
• Multi-channel and 4D image processing. The software handles time series (3D+time) and data with multiple contrasts or modalities. Segmentation tools can be applied to entire volumes or to temporal subsets.
• Manual and automatic image registration. Includes rigid and affine registration algorithms based on mutual information or cross-correlation. Registration results can be applied to align functional with anatomical images or to fuse acquisitions from different subjects.
• Segmentation interpolation between slices. To accelerate manual work, the software can generate intermediate segmentations from contours defined on non-consecutive slices. The interpolation uses basic mathematical morphology and shape models.
• 3D cut-plane tool. Allows cropping the three-dimensional reconstruction of the segmentation to visualize the interior of hollow structures or to remove unwanted regions. The cut-plane is interactive and can be freely oriented.
• Distributed segmentation service (DSS). Provides access to advanced algorithms (such as machine learning-based classifiers) running on remote servers. Users can submit images and receive segmentation results without requiring specialized hardware.
• Integrated documentation and tutorials. The program includes a quick start guide, sample data (such as the prostate reference image or calibration phantom), and links to video tutorials covering topics from installation to advanced segmentation techniques.

The development of ITK-SNAP began in 1999 at the University of North Carolina at Chapel Hill under the direction of Professor Guido Gerig. The project arose from the need for a didactic and functional tool for medical image segmentation, combining the Insight Toolkit (ITK) library for processing algorithms with a user-friendly graphical interface. Early development teams included students Joshua Cates, Tim Johnson, and Richard Goble. In 2005, Paul Yushkevich assumed project leadership at the University of Pennsylvania, incorporating substantial improvements to the interface and ITK integration. The algorithmic core is written in C++ on top of ITK, while the graphical interface has migrated from FLTK to Qt over successive versions. Current lead developers include Paul Yushkevich, Jilei Hao, Alison Pouch, and Sadhana Ravikumar, who maintain the software with contributions from the international community. ITK-SNAP remains a reference tool in neuroimaging, oncology, and biomechanics due to its balance between power and functional specialization.

Alternatives to ITK-SNAP: