ABSTRACT

In image processing, it is difficult to detect the abnormalities in brain especially in MRI brain images. Also the tumor segmentation from MRI image data is an important; however it is time consuming while carried out by medical specialists. A lot of methods have been proposed to solve MR images problems, quite difficult to develop an automated recognition system which could process on a large information of patient and provide a correct estimation. Hence enhanced k-means and fuzzy c-means with firefly algorithm for a segmentation of brain magnetic resonance images were developed. This algorithm is based on maximum measure of the distance function which is found for cluster center detection process using the Mahalanobis concept. Particularly the firefly algorithm is implemented to optimize the Fuzzy C-means membership function for better accuracy segmentation process. At the same time the convergence criteria is fixed for the efficient clustering method. The Firefly algorithm parameters are set fixed and they do not adjust by the time. As well Firefly algorithm does not memorize any history of better situation for each firefly and this reasons they travel in any case of it, and they miss their situations. So there is a need of better algorithm that could provide even better solution than the firefly algorithm. To attain this requirement as a proposed work the Artificial Fish Swarm Algorithm to optimize the fuzzy membership function. During surveying of the previous literature, it has been found out that no work has been done in segmentation of brain tumor using AFSA based clustering. In AFSA, artificial fishes for next movement act completely independent from past and next movement is just related to current position of artificial fish and its other companions which lead to select best initial centers for the MRI brain tumor segmentation. Experimental results show that presented method has an acceptable performance than the previous method.