Research Of3D Magnetic Resonance Image Brain Segmentation

In recent decades, many fields of neuroscience are affected by medical image greatly.With the rapid development of medical imaging technology, a lot of neuroscience researchin the differences of comparative brain tissue anatomy structure, so seek morphologicalchanges related characteristics of brain disease anatomical structure, so as to enhance theeffectiveness of treatment and the reliability of the brain disease diagnosis. As an importantbranch in the field of image segmentation, medical image segmentation is the first and keystep of realize the medical image analysis, and then complete medical image understandingand image analysis.With the development of Magnetic Resonance Imaging (Magnetic Resonance Imaging,MRI) technology, Magnetic Resonance (MR) Imaging can provide high resolution andhigh contrast three-dimensional (3D) medical images of brain tissue anatomy structure.The three main organization of the brain are white matter (WM), gray matter (GM) andcerebrospinal fluid (CSF), the brain is divided into subcortical structure, layer structure andpathological tissue. Neuroscience researchers gradually become strong interested in themethod that brain can be precisely segment to the above tissues. The above research basedon anatomy morphology changes depend on the understanding and segmentation ofmagnetic resonance imaging. And the medical image segmentation technology containsthese methods of extract and segment the brain tissue structure from the multimodalmedical image.However, medical image segmentation is one of the most challenging and difficultproblems in medical image analysis. Due to the limitation of magnetic resonance imagingequipment capacity, clinical collected magnetic resonance images of brain tissue usuallycontains adverse factors, such as partial volume effect, motion artifacts and uneven grayscale which caused by noise and bias field, combined with the brain tissue’s complexboundary, shape and topology structure, it makes segmentation of brain magneticresonance images accurate, fast and robust is a very difficult thing. In addition, the two-dimensional medical image segmentation has can’t meet theneeds of clinical and research,3d medical images segmentation gradually becomemainstream, clinicians and researchers are urgent toanaccurate, fast and robust3dsegmentation algorithm. Because human body organization structure is thethree-dimensional structure, and3D segmentation make full use of the3D image datainformation collected by modern imaging equipment, the segmentation result is moreaccurate and continuous in space, providing researchers with more information, such asthree-dimensional morphology, size, location, display of human tissue and the display ismore intuitive and clear.According to the modern development of medical image segmentation field, in theview of3d MRI brain image segmentation, this article analysis and review the currentmajor medical image segmentation methods detailedly, and especially the3d medicalimage segmentation algorithm put forward a new3d MRI brain image segmentationalgorithm, and using the algorithm can succeed to segment three-dimensional tissue, andcomplete the research content as following several aspects:Firstly, because of the existing of the noisy and Bias field for brain resonance image, Itake advantage of Gaussian filter to deal with Three-dimensional data for its denoising atfirst, to finish the denoising and smooth of that brains resonance imaging which containsthe noisy and Bias field, and make ready for the latter segmentation.Secondly, it is come true for the Image algorithms on brain magnetic resonanceimage which based on the Threshold segmentation and Regional growth in ITK, andarguments that the effectiveness to brain issue MRI segmentation, which has laid thefoundation of the theory and practicefor the3D medical image segmentation.Thirdly, it has final treatment for segmentation result on the basis of those contents, whichtake advantage of Mathematical morphology and Image "and" operation, and makesegmentation result more accurate Close to the Real anatomical structure of the brain issue.On the basis of the segmentations above, I have introduced the existing Segmentationevaluation method in detail. The experimental result shows that this paper the selected ITK algorithm has winedthe ideal segmentation results, which keep off the phenomenon for segmentationincorrectly of the Gray matter and CSF by its easy algorithms, fast segmentation speed andstrong robustness, what have offered a good foundation of the study of the internalcharacteristics of brain tissue furthermore.