Brain MRI Segmentation using Canny Edge Detection Technique

Abstract

Image segmentation is a very tough technical process in the field of image processing. It is mandatory to go for the pre-processing before actual identification and segregation of ROI in MRI. Edge detection is based on the identification of the bounadries on the basis of similar brightness level and continuity of some pattern in the pixels. It is not possible to separate the some segment without applying the pre-processing of the image. . The applications of medical image segmentation are 3D reconstruction and quantitative analysis. Our experiment uses the MRI because this type of images give the most reliable image of the internal human organs like brain.. In this paper, the common problems of the image segmentation are explored and a novel method of image detection has been used.The classical image segmentation methods have been discussed for example gradient based operators were used usedfor edge detection, but these methods could deliver in the noisy surroundings. In the medical applications, the precision cannot be compromised with. To overcome these difficulties, we have proposed Canny edge detection technique prior to NonLocal Fuzzy C-Means Clustering technique for the segmentation. Quantifying brain structures in such large databases cannot be practically accomplished by expert neuroanatomists using hand-tracing. Rather, this research will depend upon automated methods that reliably and accurately segment and quantify large number of brain regions. At present, there is little guidance available to help clinical research groups in choosing such tools. This work is targeted to find out and establish more reliable segmentation technique as compared to expert hand tracing. The proposed approach consistently gives better results for various noise levels in the image compared to the reference schemes. This method is used for detecting brain region based on their energy function. In order to compare between them, one slice of MRI image tested with these methods. The traditional and proposed edge detection algorithms are implemented in MATLAB and results of proposed method are presented and compared with traditional approach. [1]