Biological Early Detection For Brain Cancer Using Artificial Neural Network

Abstract

This thesis aims to study how the artificial neural network which is one of the artificial intelligence (AI)techniques inthe application ofbiomedical images identified and determining its kind, also has demonstrated the idea of artificial neural networks in the vital areas of research related to the medicine field, because ANN are currently a hot researches in medical area particularly in field of radiology, cardiology, oncology etc. Also the research represents a practical model for the use of neural networks to identify the medical images, which is considered as anearly detection system of the brain tumor. Neural Network must able to determine the state of brain tumor according to magnetic resonance imaging and determine is it normal or abnormal from each MR image, Harlick texture feature will extract to prepare training data which was introduced to neural network as inputs and target vector, two network was designed and trained using MATLAB feature nntoolwhich are Feed forward back propagation and Cascade-forward back propagation, after testing cascade network achieved performance ratio 86.6% and FFBP achieved 91.42%.