Segmentation Of 3D MRI Image Of Children Brain

MRI (Magnetic Resonance Imaging) has been a very significant part of medical imaging techniques, which is recognized as a very important tool for clinical diagnosis. The main task in MRI is to infer useful information from the images for diagnostic purposes. MRI proved to provide high quality medical images and became widely used especially for brain. Today, many current problems in image-guided surgery, therapy evaluation and diagnostic tools strongly benefit from accurate 3D models of anatomical structures. This implies that the study of segmentation methods is the most important part of all the research. Segmentation of the MRI images involves the isolation of anatomical structures from images obtained, considered as the primary goal of providing accurate representations of key anatomical structures. The visualization and quantification of the segmented images can provide an anatomical framework of brain and help us to do the brain functional research. Moreover, the use of segmented brain image is widely applied in cortical surface mapping, volume measurement, tissue classification and differentiation, etc. Up to now, lots of studies have addressed to the segmentation of adult's brain, but few of them has concerned the case of children's. In this paper, We study a semi-automatic framework of segmentation of children's brain MRI image ( T1 weighted), which is capable of identifying the different structures of children's MRI images by using the histogram analysis and morphological operations. The framework consists of four-step segmentation procedures. First, the non-brain structures removal is addressed to obtain the mask of encephalon, then separate the brain stem and cerebellum respectively from the encephalon mask, and finally, two hemispheres are separated. The method gives the good segmentation results in the children MRI image aging from 5-15 years old.