MeVisLab is a cross-platform application framework for medical image processing and scientific visualization. It includes advanced algorithms for image registration, segmentation, and quantitative morphological and functional image analysis. An IDE for graphical programming and rapid user interface prototyping is available. MeVisLab is written in C++ and uses the Qt framework for graphical user interfaces. It is available cross-platform on Windows, Linux, and Mac OS X. The software development is done in cooperation between MeVis Medical Solutions AG and Fraunhofer MEVIS. A freeware version of the MeVislab SDK is available (see Licensing). Open source modules are delivered as MeVisLab Public Sources in the SDK and available from the MeVisLab Community and Community Sources project.
History MeVisLab development began in 1993 with the software ILAB1 of the CeVis Institute, written in C++. It allowed to interactively connect algorithms of the Image Vision Library (IL) on Silicon Graphics (SGI) to form image processing networks. In 1995, the newly founded MeVis Research GmbH (which became Fraunhofer MEVIS in 2009) took over the ILAB development and released ILAB2 and ILAB3. OpenInventor and Tcl scripting was integrated but both programs were still running on SGI only. In 2000, ILAB4 was released with the core rewritten in Objective-C for Windows. For being able to move away from the SGI platform, the Image Vision Library was substituted by the platform-independent, inhouse-developed MeVis Image Processing Library (ML). In 2002, the code was adapted to work on the application framework Qt. In 2004, the software was released under the name MeVisLab. It contained an improved IDE and was available on Windows and Linux. See the Release history for details. In 2007, MeVisLab has been acquired by MeVis Medical Solutions AG. Since then, MeVisLab has been continued as a collaborative project between the MeVis Medical Solutions and Fraunhofer MEVIS.
Features
MeVisLab features include:
Image processing with the MeVis Image Processing Library (ML): The ML is a request-driven, page-based, modular, expandable C++ image processing library supporting up to six image dimensions (x, y, z, color, time, user dimensions). It offers a priority-controlled page cache and high performance for large data sets. 2D image viewing: Fast, modular, extensible 2D viewers with combined 2D/3D rendering are implemented, supporting slab rendering (volume rendering/MIP), overlays, point/ROI selection, Multiplanar Reformations (MPR), as well as interactive editing of marker objects (points, vectors, discs, spheres, etc.) Volume rendering: A high-quality volume renderer (Giga Voxel Renderer, GVR) based on OpenGL/Open Inventor is available. It supports large image volumes (e.g., 512x512x2000 CT volumes, 12bit), time-varying data (e.g. dynamic MRI volumes), lookup tables, interactive region of interest, sub-volume selection, modular, multi-purpose GLSL shader framework. DICOM and other file formats: DICOM is supported via an import step that automatically recognizes series of 2D DICOM frames that belong to the same 3D/4D image volume. The data can be browsed with a configurable DICOM browser. DICOM storage to PACS is possible. Other supported file formats include TIFF (2D/3D, RGBA), Analyze, RAW, PNG, JPG, BMP, and more. Tool frameworks: Modular class and module libraries for markers, curves, histograms, Winged-Edged Meshes (WEM) and Contour Segmentation Objects (CSO) are available. Qt integration: Qt is used as application framework. The Qt API is integrated via PythonQt, allow to access Qt Style Sheets, Qt Widgets, QT Core classes, etc. by scripting from within MeVisLab. Scripting support: Python can be used for script controlled access to a large part of the MeVisLab functionality. The script binding to Qt is implemented via PythonQt. For image processing via Python, NumPy is available. Object-oriented Python programming in MeVisLab is possible. Integrated open source image processing and visualization libraries: Three open source libraries are integrated: Open Inventor, based on the original SGI source code released as open source in 2000; Insight Toolkit (ITK), made available as MeVisLab modules; Visualization Toolkit (VTK): made available as MeVisLab modules. Comprehensive module library: The MeVisLab module library comprises a total of 2600 modules, including 800 standard modules and 1800 ITK/VTK modules.
MeVisLab principles
MeVisLab is a modular development framework. Based on modules, networks can be created and applications can be built. To support the creation of image processing networks, MeVisLab offers an IDE that allows data-flow modelling by visual programming. Important IDE features are the multiple document interface (MDI), module and connection inspectors with docking ability, advanced search, scripting and debugging consoles, movie and screenshot generation and galleries, module testing and error handling support. In the visual network editor, modules can be added and combined to set up data flow and parameter synchronization. The resulting networks can be modified dynamically by scripts at runtime. Macro modules can be created to encapsulate subnetworks of modules, scripting functionality and high-level algorithms. On top of the networks, the medical application level with viewers and UI panels can be added. Panels are written in the MeVisLab Definition Language (MDL), can be scripted with Python or JavaScript and styled using MeVisLab-internal mechanisms or Qt features. The development of own modules written in C++ or Python is supported by wizards.
Image gallery
MeVisLab forum MeVisLab offers a forum which requires registration to use. Features include, BBS support, private messaging, and a calendar.
Fields of application, research projects
… excerpt ends here. Continue reading the full article.






