Brain Tumor Detection that uses CNN in MRI
The experimental results of the proposed classification method based on Convolutional Neural Network for MR images shows that it have good accuracy and is productive and have very low intricacy rate and comparitevely less execution time.
Abstract
Brain cyst is a mass of unnecessary units growing in the brain or the multiplication of normal or abnormal cells that are not essential for the brain, which may even be the cause for death. The brain tumors diagnosis is the initial step of medication which is a time consuming and delineating process by expert radiologist and depends on interrater reliability. Magnetic Resonance Image of brain is an important factor in primitive surgical preparations and in computer aided surgery. In order to augment the effort of radiologists in retrieving the information from MRI, we use this system. This proposed paper have classification method based on Convolutional Neural Network for MR images.The dataset is taken from Kaggle and is augmented to get a bigger dataset and to avoid overfitting; consequently preprocessing of MRI and proposed model compilation is done. It also compares the proposed architecture with the VGG16 architecture to compare the outcome and advantages. The experimental results of the model shows that it have good accuracy and is productive and have very low intricacy rate and comparitevely less execution time. Performance comparison is also done with the classification report and the confusion matrix. The VGG16 model acquired an accuracy of 59% and the proposed model acquired an accuracy of 91.2% with a good accuracy.