The Method For Segmentation Of Brain Tissues In MRI

As a new imaging technique, MRI plays an important role in the research of brain function field, because its non-invasion and clear imaging. Consequently, with the need of 3D reconstruction, quantitative analysis and visualization, the segmentation of head MRI is required to be more precise and fast. However, the segmentation of MRI would be more difficult due to the following factors: 1) indistinct boundaries between brain tissues. 2) variance in brain structures among individual. 3) intrinsic uncertainty of MR images and induced heterogeneity of magnetic field. Therefore, the segmentation of head MRI becomes a focus issue in the medical image research. Consequently this paper intends to do more research in the segmentation of head MRI with neural network, fuzzy logic and anisotropic diffusion.The anisotropic diffusion and mean shift algorithm is designed to exactly segment the brain structure such as the white matter, gray matter and the cerebrospinal fluid. Fuzzy logic algorithm is common segmentation methods to MRI, we combine this method with the neural networks and obtain the satisfied result. In the process of the experiments, we conclude and suggest for the future work in head MRI segmentation field.