Multiclass Brain Tumor Classification Using Transfer Learning

Tumors are a collection of abnormal cells that multiply enormously than required which leads to cancer and divergent and also can be fatal, if not identified at an early stage. Usually, brain scan described as Magnetic resonance imaging (MRI) is deployed for high transparency and representation in different angles but causes huge delay in declaring the result of the test. In this project, images obtained from these tests are carefully observed and classified by implementing Deep Residual Network (RESNET) to classify the type of tumor. There are four types of tumors such as glioma, meningioma, pituitary, and no tumor. Brain tumor classification (Multi Label) – CNN dataset has been imported to train and test the model. This deep learning model is a sophisticated approach which is developed to classify the tumor based on the image, so that appropriate treatment can be given on time. The output determines the type of tumor if present, otherwise no tumor with accuracy of 87% using epochs.