a comparative study on tissue classification of brain mr images using dipy, spm, and fsl frameworks

Introduction: image classification is a disputable part of image processing, especially brain tissue classification in magnetic resonance images because brain tissue signals and contrasts are close to each other. more accurate classification of mr images of the brain is crucial for the diagnosis of csf, gray, and white matter. material and methods: a bayesian model was used to classify 20 brain mr images using the criteria of dipy, spm, and fsl. two distinct dice and jaccard coefficients were used to assess the similarity of all classified images. results: spm classification was more accurate than dipy and fsl in categorizing cerebrospinal fluid. the dice and jaccard coefficients for the spm classification were 97.48 ± 0.28 and 92.68 ± 0.94, respectively. the dice and jaccard coefficients for white matter were 95.64 ± 0.23 and 86.18 ± 1.64, respectively, while the coefficients for gray matter were 93.66 ± 0.76 and 83.62 ± 1 .92 for the dipy. conclusion: the dipy python library was able to better cluster gm and wm regions according to the results obtained.