MRI Brain Image Segmentation Based On Improved Energy Minimization Framework Combined With SVPSO-NKFEC Algorithm

"Stroke",also known as "stroke",is a kind of cerebrovascular disease caused by the sudden rupture of blood vessels in the brain or blocking blood flow to the brain,which will eventually lead to brain tissue damage,high disability rate and high mortality rate.MRI technology is a nondestructive imaging,and can obtain high resolution images of tissue structure,accurate segmentation of brain tissue is the premise of stroke diagnosis of the degree of brain tissue injury.MRI brain image segmentation method is commonly used in nuclear fuzzy entropy clustering method,while the algorithm by optimizing the membership value and high dimension of input data are mapping,eliminates some noise in the MRI brain images,but the three main problems: for not considering the image neighborhood spatial information,which results in the image segmentation is still affected by the noise;Because of the sensitivity to the initial value,the clustering solution is easy to fall into local optimum.Because of the inhomogeneity of the offset field,the brain tissue strength range overlaps and leads to missegmentation.Aiming at the deficiency of KFEC ignoring spatial information,this paper proposes a kernel fuzzy entropy clustering algorithm based on spatial information,NKFEC for short.This algorithm takes advantage of Gaussian kernel fuzzy entropy clustering to have strong robustness to outliers or noise of data,and adds neighborhood spatial information,so that the new algorithm retains the feature consistency of neighborhood pixels.In view of the problem that initialization is easy to fall into local optimal,the particle swarm optimization algorithm is initialized and inertial weight is improved,called SVPSO,and NKFEC is introduced to propose a segmentation algorithm based on the combination of improved particle swarm optimization and NKFEC algorithm,abbreviated as SVPSO-NKFEC.In view of the inhomogeneity of the migration field,the energy minimization framework was improved in this paper,and a deviation correction method called BE was proposed by integrating the migration field and the ideal image properties into the framework,and a segmentation algorithm based on the combination of the improved energy minimization framework and the SVPSO-NKFEC algorithm was introduced,which was called BE-SVPSO-NKFEC.The experimental results show that the NKFEC algorithm takes into account the neighborhood spatial information between pixels in the segmentation process,eliminates the influence of noise on pixels to the maximum extent,and improves the segmentation accuracy of MRI brain images.SVPSO-NKFEC algorithm uses the improved particle swarm optimization to optimize the clustering solution of NKFEC algorithm,which can search the space more effectively at the beginning of iteration,avoid falling into the local optimal solution,and improve the accuracy and speed of MRI brain image segmentation at the same time.The BE-SVPSO-NKFEC algorithm uses the improved energy minimization framework to estimate the offset field,which effectively overcomes the effect of the inhomogeneity of the offset field on the segmentation,so that the algorithm has a strong ability to resist the offset field and can obtain a high accuracy of MRI brain image segmentation.