Implementing Texture Analysis Software Frame for Magnetic Resonance Image Data in MATLAB
Varjo, Jukka (2014)
Varjo, Jukka
2014
Sähkötekniikan koulutusohjelma
Tieto- ja sähkötekniikan tiedekunta - Faculty of Computing and Electrical Engineering
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2014-06-04
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tty-201406061272
https://urn.fi/URN:NBN:fi:tty-201406061272
Tiivistelmä
This thesis was based on the need to develop a generic software application frame for texture analysis of magnetic resonance (MR) images. In collaboration with the research group at the department of Medical Imaging Centre and Hospital Pharmacy (MICHP) at Tampere University Hospital (TAUH) the goal was to improve the user experience and work flow as well as implement a completely new user interface and key functionalities. The platform was required to be complex enough to manage with image processing algorithms and to provide high level and easily modifiable software architecture. The research group having years of experience with an open-source texture analysis oriented MaZda software the focus of this thesis was to analyse and solve the restrictions based on the observations from using MaZda.
MATLAB was chosen as the programming platform due the high-level syntax with powerful built-in properties e.g. Image Processing Toolbox (IPT) that would allow proficient support for computationally demanding processes. Another advantage with MATLAB was the interface support for languages like Fortran, C and C++. MATLAB being commercial software platform, it was acknowledged that achieving a standalone end product would not be possible. Computational performance was also omitted for the purpose this thesis not only due to MATLAB’s limitations but also to keep the scale contained.
The improvement suggestions provided by the research group were considered as a rough specification for the software to be implemented. These requirements included extensibility in terms of texture analysis algorithms and simplified user interface to improve the work flow. Selecting MATLAB as the programming environment extended the group of people capable of contributing to the tool in the future. Implementing the frame from the beginning allowed the texture analysis parameters and features to be fully configurable instead of static. The modular visual structure of the software allowed the user to switch between image sets more easily. Removing the region of interest (ROI) limitation ensured that same image set could be utilized more efficiently.
The implemented MATLAB application provides a basic frame for more convenient medical image processing flow for texture analysis of MR images but further testing and development is required to complement the tool.
MATLAB was chosen as the programming platform due the high-level syntax with powerful built-in properties e.g. Image Processing Toolbox (IPT) that would allow proficient support for computationally demanding processes. Another advantage with MATLAB was the interface support for languages like Fortran, C and C++. MATLAB being commercial software platform, it was acknowledged that achieving a standalone end product would not be possible. Computational performance was also omitted for the purpose this thesis not only due to MATLAB’s limitations but also to keep the scale contained.
The improvement suggestions provided by the research group were considered as a rough specification for the software to be implemented. These requirements included extensibility in terms of texture analysis algorithms and simplified user interface to improve the work flow. Selecting MATLAB as the programming environment extended the group of people capable of contributing to the tool in the future. Implementing the frame from the beginning allowed the texture analysis parameters and features to be fully configurable instead of static. The modular visual structure of the software allowed the user to switch between image sets more easily. Removing the region of interest (ROI) limitation ensured that same image set could be utilized more efficiently.
The implemented MATLAB application provides a basic frame for more convenient medical image processing flow for texture analysis of MR images but further testing and development is required to complement the tool.