Adaptive glioblastoma detection using evolutionary-based algorithm / Nurul Amira Mohd Ali

Brain tumor is a common disease with low survival rate. Various types of brain tumor are exist but Glioblastoma is the most aggressive and dangerous among all of it which has caused plenty of deaths. It required an early diagnosis to prevent it from spreading therefore, the brain tumor need to be detected early. The process of brain tumor detection usually done by a radiologist, but it may be susceptible to errors due to large amount of patients and any late diagnosis may harm the patient. Hence, this project was proposed to help in overcome the problems. The objectives of the project are to design and develop a prototype of adaptive Glioblastoma detection using Evolutionary-based algorithm to assist in detecting brain tumor and also to test the prototype’s functionality and detection accuracy. Artificial Bee Colony algorithm has been selected as the algorithm used in segmenting the brain tumor from MRI image. It was inspired by honeybee’s foraging behavior. The prototype’s functionality testing proves that the prototype works well, and the detection accuracy testing proves that the algorithm shows a good performance in producing the output of segmented image. The testing on adaptive glioblastoma detection produced 93.51% of average accuracy indicates the prototype produced a good glioblastoma segmentation result by using ABC algorithm. In conclusion, all three objectives have been achieved as the glioblastoma detection prototype has been designed, developed and tested.