The primary notion relying in image processing is image segmentation and classification. The intention behind the processing is to originate the image into regions. Variation formulations that effect in valuable algorithms comprise the essential attributes of its region and boundaries. Works have been carried out both in continuous and discrete formulations, though discrete version of image segmentation does not approximate continuous formulation. An existing work presented unsupervised graph cut method for image processing which leads to segmentation inaccuracy and less flexibility. To enhance the process, our first work describes the process of formation of kernel for the medical images by performing the deviation of mapped image data within the scope of each region. But the segmentation of image is not so effective based on the regions present in the given medical image. To overcome the issue, we implement a Bayesian classifier as our second work to classify the image effectively. The segmented image classification is done based on its classes and processes using Bayesian classifiers. With the classified image, it is necessary to identify the objects present in the image. For that, in this work, we exploit the use of pattern matching algorithm to identify the feature space of the objects in the classified image. An experimental evaluation is carried out to estimate the performance of the proposed efficient pattern matching algorithm for classified brain image system [EPMACB] in terms of estimation of object position, efficiency and compared the results with an existing multi-region classifier method.

Image segmentation Classification Pattern matching similarity measure