期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2015
卷号：15
期号：11
页码：21-26
出版社：International Journal of Computer Science and Network Security
摘要：Magnetic Resonance Imaging (MRI) brain images often suffer from low contrast and noise, especially in brain imaging. This noise hampers further tasks such as segmentation of the important features and classification of brain tumor. As a result, the visual quality gets deteriorated and perfect diagnosis of the disease becomes difficult. During the acquisition process of (MRI), irregular bias is imposed in the intensity values of the pixels. These biases follow the Gaussian Noise distribution model and act as a constraint to the effective medical diagnosis. We are interested in Partial Differential Equations (PDE) in order to smooth MRI brain image in an anisotropic manner. The Anisotropic Diffusion filter (ADF) approach is limited to preserve the structural integrity of MRI brain image at only low noise levels. This paper proposes (CADF) algorithm aimed to improve the estimation of the diffusion constant to facilitate better edge detection and preservation of details. We demonstrated how the diffusion tensor computed in an anatomically enhanced MRI brain image coordinate by (CADF).This framework facilitates radiologist to assess brain tissue change and guide them to evaluate MRI brain image of having brain tumor Simulation trials have been conducted at different Gaussian noise variances and performance has been evaluated on the basis of Peak Signal- Noise Ratio (PSNR) and Structural Similarity (SS). The proposed algorithm has shown stable value of evaluation parameters at higher noise variances. Also, the preservation of details has improved as compared to the Curvature Anisotropic Diffusion Filter. Finally, we illustrate the efficiency of our generic curvature-preserving approach. Experimental results show that the new method can achieve better denoising results in a variety of MRI brain images, and the new approach shows superior performance on edge and curvature preserving edges and texture image.
