Artificial Bee Colony Algorithm for Adaptive Glioblastoma Detection

Abstract

Glioblastoma is a most dangerous and aggressive high-grade brain tumour. The high-grade tumors necessitate early detection and treatment due to the rapid growth rate, and early diagnosis may improve the chance of survival. The glioblastoma detection is currently done by a radiologist, however it is time-consuming, invasive, and prone to errors due to the enormous volume of cases. Thus, in this study, the Artificial Bee Colony (ABC) algorithm was employed to provide a non-invasive approach of adaptive glioblastoma detection. The feature properties of the glioblastoma were studied using the basic feature analysis of Minimum, Maximum, and Mean of grey level values. Four different types of T1-weighted, T2-weighted, Fluid Attenuated Inversion Recovery (FLAIR), and T1-contrast MRI images were used to assess the ABC's performance for adaptive glioblastoma detection. A total of 120 MRI glioblastoma images were evaluated, with 30 images per imaging category. The overall mean percentage of accuracy for glioblastoma detection was 93.67%, indicating that the suggested adaptive ABC algorithm has a high capability for glioblastoma brain tumor detection. Other feature extraction strategies, however, could be introduced in the future to improve the feature extraction performance.