Segmentation for Medical Image Mining: A Technical Report

Magnetic resonance imaging (MRI) of the brain, followed by automated segmentation of the Corpus Callosum in midsagittal sections has important applications in neurology and neurocognitive research since the size and shape of the Corpus Callosum are shown to be correlated to sex, age, neurodegenerative diseases and various lateralized behaviour in people. The segmenting of the Corpus Callosum is regarded as a critical step in image mining frameworks to classify the brain MR images. In the example presented here a test collection of 76 brain MRI images representing musician and non-musician is used. Two algorithms, based on established work, for segmenting the Corpus Callosum in brain MR images, are presented and evaluated; based on this evaluation a new algorithm is also proposed. The segmentation algorithm operates by first extracting regions satisfying the statistical characteristics (gray level distributions) of the Corpus Callosum that have relatively high intensity values. This is then processed using graph analysis and classification procedures. Test using the musician data set have provided promising results.