The MITK Approach
Abstract
The Medical Imaging Interaction Toolkit (MITK) is an opensource toolkit for the development of interactive medical image analysis software. MITK is based on the open-source Insight Toolkit (ITK) and Visualization Toolkit (VTK) and extends them with features required for interactive systems. ITK is used for the algorithmic scope and general infrastructure, VTK for visualization. Key features of MITK are the coordination of multiple 2D and 3D visualizations of arbitrary data, a general interaction concept including undo/redo, and its extendibility and flexibility to create tailored applications due to its toolkit character and different layers of hidden complexity. The paper gives a brief introduction into the overall concepts and goals of the MITK approach. Suggestions and participation are welcome. MITK is available at www.mitk.org.
Keywords
Source Code and Data
No source code files available for this publication.
Reviews
Stephen Aylward
Monday 19 September 2005
Summary: The authors present MITK, a freely available and truly open-source (commercial and non-commercial use) software library that provides modules that integrate and extend ITK, VTK, GLUT, and FLTK/QT. Integration alone is noteworthy - the extensions are even more noteworthy.
Key features include the coordination fo multiple 2D/3D visualizations of spatial objects, a general interaction concept including undo/redo, and a layered API that hides complexity at upper levels while also allowing specialization of functionality at lower levels.
The layered approach is particularly useful in this context. A user can quickly piece together some broad modules to form a “standard” application, and then time can be spent to refine the inner workings of one or two of the modules to provide tailored functionality.
Hypothesis: A diversity of open-source projects can be integrated and wrapped to provide software modules that simplify medical image analysis application development.
Evidence: The paper is written in a casual tone and introduces the concepts behind MITK and provides a few specifcs regarding the options available in MITK.
The paper provides an excellent overview of which strengths were combined from the various open source packages that comprise MITK. That overview is written at an ideal level of detail. Specific programming constructs (e.g., smart pointers) are mentioned without going into overly detailed explanations. The details given are exactly what a reader would need to begin to investigate the integration of those packages into his/her own toolkit.
The paper, however, would benefit from a bit more details regarding what options are available to someone using MITK to develop an application. A list of available visualization models, for example. Otherwise, the authors could provide more details regarding the process of developing one or two demonstration applications using MITK. Either type of additional information would be welcome.
Open Science: MITK is released under an open-BSD-style license. It is available for commercial or non-commercial use. Many aspects are GUI independent, and GUI components from different packages (e.g., QT or FLTK) can be used; however, most work appears to be done with QT components - the development of FLTK components seems to be more limited. This is unfortunate given QT’s licensing restrictions.
Reproducibility: The website for MITK is very good. Publications, documentation, and the code are readily available. The installation details and requirements are well specificied.
Use of Open Source Software: Makes heavy use of open-source.
Open Source Contributions: Is released as open source: commercial and non-commercial.
Code Quality: I did not look at the code, but the use of ITK, VTK, etc and the fact that the group has contributed back to those toolkits suggests that they use good coding style/practices.
Additional Comments: Users interesting in this package should be aware of related work in the form of Slicer and SOViewers. MITK, however, appears to fill a gap between those two toolkits. It allows more specialization than Slicer, and it is easier to develop a standard application using MITK than using SOViewers.
David Holmes
Tuesday 16 August 2005
Summary: This paper describes the development of a medical image software framework based on ITK, VTK, QT, and FLTK. The framework is, conceptionally, one level closer to the end-user than any of the components, because it creates high-level widgets and controls used for interacting with data and algorithms. End-user developers should be able to build interactive applications with MITK.
Hypothesis: Not Applicable
Evidence: Not Applicable other than the fact that the framework has been built. The paper doesn’t provide studies or evidence that the framework is valid, end-user acceptable, fast, or other metrics of performace. However, this was not the intent of the paper and (my opinion only) I don’t believe such a study would be prudent at this time. (my opinion only) I would suggest that the software probably requires some time to mature.
Open Science: The work is consistent with open science principles
Reproducibility: The software can be downloaded and run. Although the package hasn’t been submitted with the paper, the paper explicitly describes where the code can be obtained from. I admit that I have not taken the time to download and run it.
Use of Open Source Software: Most of the packages are open source except for QT which does have some licensing restrictions.
Open Source Contributions: The whole package is available for user evaluation. In addition, the authors request community participation to test and develop the software.
Code Quality: Once again, in the intereste of full disclosure, I have not yet downloaded the package and worked with it.
Applicability to other problems: Because it is a framework, there are many specific applications which can utilize this code to develop targeted software.
Suggestions for future work : Performace evaluation in the context of specifica applications.
Requests for additional information from authors : It would be beneficial to have a little code example in the text to show the implementation approach. As is noted in figure 2, a small example in fltk exists. Some amount of this code might be included and described in the text as a specific example.
Additional Comments: None