Research Of Medical Image Segmentation And Bias Field Correction Based On Fuzzy Clustering

With the rapid development of medical imaging technology,medical imaging has gradually become one of the main auxiliary methods in clinical medical,and it improves the doctor's diagnosis rate to a large extent.Currently,medical image segmentation techniques focus on magnetic resonance imaging(MRI).Brain tissues include gray matter(GM),white matter(WM)and cerebro-spinal fluid(CSF).Accurate brain tissue segmentation from brain magnetic resonance(MR)image is a critical step in quantitative brain image analysis,it's also a key and difficult issue in the field of medical image processing.As the brain MR image is influenced by partial volume effect,noise and intensity inhomogeneity,the existing segmentation algorithms are more or less incorrect.This is the key issues and difficulties on medical image processing,so brain medical image segmentation has caused a wide range of researchers' attentions.Fuzzy C-Means(FCM)is the most widely used fuzzy clustering image segmentation algorithm.Although the improved FCM algorithms has segmented different types of images successfully,they have the following weaknesses in segmenting brain MR images.First,they can not correct bias field effectively on those MR images with bias field,resulting in incomplete segmentation results;Second,the degree of influence on the neighborhood pixels are not accurate,resulting in algorithms are not ideal for noise immunity.Based on the analysis above,this paper presents a new fuzzy c-means algorithm(RCLFCM)for segmentation and bias field correction of brain MR image.First,considering that the FLICM algorithm only considers the spatial distance factor,the influence of the neighborhood window can not be estimated correctly.To solve this problem,RCLFCM algorithm presents a new neighborhood gray-difference coefficient and combines it with local window variance coefficients,designs a new influence factor to measure the effect of neighborhood pixels,so the robustness of anti-noise can be enhanced.Besides,this paper constructs a new spatial function by combining pixel gray value similarity with its membership,and make full use of the space information between pixels to update the membership.So that the algorithm can achieve convergence in less number of iterations and improve the effectiveness of the algorithm.Finally,this paper redefines the objective function of FCM by adding the bias field estimation model to overcome the intensity inhomogeneity in the image and segment the brain MR images simultaneously,which suppresses the influence of the bias field on the medical image and improves the segmentation accuracy.In this paper,the new algorithm is used to synthesize brain images of MRI with bias and noise,and the experimental results are compared with five other representative algorithms.After the visual effect analysis and the segmentation accuracy analysis,the experimental results show that the proposed algorithm can estimate the bias field and suppress the noise more effectively,and obtain more accurate segmentation results of brain tissues.