Brain tumor classification using modified genetic algorithm based FCM and deep belief network
Abstract
Brain tumor is one of the most dangerous disease in the world and it is caused by the abnormal growth of cells in the brain. The segmentation and classification of the brain tumor is important to extract and classify the type of tumor. The classification of tumor type is helpful in diagnosis and treatment planning. In this research paper, an effective brain tumor segmentation and classification technique is developed to segment and classify the brain tumor. The modified Genetic Algorithm (GA) based Fuzzy C-Means (FCM) technique is used to segment the brain tumor by combining local and global search probability. Three different feature extraction techniques such as Speeded up Robust Feature (SURF), Binary Robust Invariant Scalable Key points (BRISK) and Local Ternary Pattern (LTP) are used to extract the features from the segmented brain tumors. The classification among different brain tumor types is achieved by using Deep Belief Network (DBN). The performance of the proposed method is validated in two datasets such as Digital Imaging and Communications in Medicine (DICOM) and BRATS. The performances are analyzed in terms of accuracy, precision, sensitivity and specificity. The performance of the proposed method is compared with two existing methods such as Run Length of Centralized Patterns with Naive Bayes (RLCP-NB) and Modified FCM with multi objective variable string length based on GA (MFCM-MOVGA). The accuracy of the high grade tumor classification for BRATS dataset is 97.59%, it is high when compared to the RLCP-NB method.