Framework for abnormality detection in magnetic resonance brain images
Aleksandar Stojak, Eva Tuba, Milan Tuba · 2016
Magnetic resonance images (MRI) are often the most powerful diagnostics tool in medicine. They are especially useful for brain examination where precise differentiation of various tissues is possible with the goal of discovering abnormalities. Different abnormalities appear in images as regions with different features. In this paper we propose a software framework for segmentation of abnormalities in brain MRI. Main algorithm is block based modified region growing segmentation where homogeneity criteria are predetermined for typical abnormalities and can easily be adjusted for changes in images parameters. Framework was tested and compared with other approaches from literature on standard benchmark brain MRI and it proved to be capable of precise and accurate segmentation of abnormal regions.