An Automated Hybrid Methodology Using Firefly Based Fuzzy Clustering for Demarcation of Tissue and Tumor Region in Magnetic Resonance Brain Images
Every year, many people die due to brain tumor. Currently, tumor region identification located in the human brain by radiologist plays a significant role. Even though many techniques have been developed for tumor region identification, but still manual assistance is required for diagnosis in magnetic resonance (MR) image. There is a need for automated methodologies for identifying the region of tumor and segmenting the tissues in magnetic resonance (MR) image. The automated segmented algorithm based on Firefly based clustering will improve the diagnosis of tumor region located in the human brain (without any manual assistance) in minimum interval of time. The Firefly algorithm is utilized to find the optimal cluster location, from which the clustering operation is performed by Interval Type-2 Fuzzy C-Means (IT2FCM). The optimal cluster location identified by the Firefly algorithm helps in easing the clustering operation performed by IT2FCM algorithm, and thereby reducing the computational complexity. The suggested techniques are compared with conventional bio-inspired algorithm and it will give better segmentation result in identifying various types of pathologies in the human brain. The potency of the proposed technique is validated by comparison parameters such as Tanimoto coefficient index, Dice overlap index, and elapsed time. Finally, it is confirmed that the recommended Firefly based IT2FCM clustering techniques produce better augmentation of demarcation of MR brain image and assists the radiologists to make accurate decision.