Data Augmentation Based Brain Tumor Detection Using CNN and Deep Learning

Abstract

Brain tumor, always being under spotlight threatening human's life expectancy. The high chances of saving people from this life-threatening disease opens the gateway to our research of brain tumor detection. Magnetic Resonance Imaging (MRI) is being more efficiently utilized to easily identify the brain tumor. As this medical diagnosis branch of brain tumor detection has a very limited exposure to having datasets with larger sizes, the abundancy to implement Data Augmentation is also equally high. Data Augmentation stands to enrich the existing dataset and the common ways of augmentation are implemented. The implementation of Machine Learning/Deep Learning algorithms in the health industry has also increased exponentially greater over the years. Convolutional neural networks (CNNs) have been resonating in the application area with their DL approach. The research work supports the implementation of models such as VGG-16, ResNet-50, DenseNet121. Based on the results we obtain an efficient model is proposed for the detection of brain tumor.