Brain Tumor Classification with Optimized Features using Firefly Algorithm

In present scenario tumor is a dangerous disease in the human life. In this study, the number of steps of medical image processing are used to process a Magnetic Resonance Image (MRI) to detect the brain tumor and also to define its type. Out of many steps, feature extraction and feature reduction is core of medical image processing. Different types of features like intensity, shapes and texture based features, are extracted from the segmented MRI images. Then, a machine learning model is developed by using the optimally selected features. Feature reduction approach is used to select the small subset of features which minimize redundancy and maximize relevance to the target. A bio-inspired Evolutionary computation (EC) based Firefly Algorithm (FA) has been used to reduce the size of feature set so that only key features will be used to classify tumor type using Support Vector Machine (SVM). At the end, classification accuracy has been compared results obtained with and without optimization.