A Hybrid Model for Brain Tumor Detection using EfficientNet and Fuzzy C Means Clustering Algorithm

Abstract

This paper describes the detection of brain tumors from Magnetic Resonance Images (MRI) using the deep learning EfficientNet model and the Fuzzy C means algorithm. The earlier detection of brain tumors can reduce the risk of death. Deep Learning is a highly adoptable technique for the detection of brain tumors at an early stage. It potentially lowers the fatality rate rather than machine learning as it allows the processing of large amounts of data to be more accurate in medical diagnosis. The existing EfficientNet model did not include a segmentation algorithm which is required for model training because it provides a clear image of the training model that is deployed in the EfficientNet model. It was reviewed that the Fuzzy C Means clustering algorithm is the best used for segmentation along with the enhancement of the existing EfficientNet model. This proposed system uses a transfer learning approach for training the model which was evaluated that led to an accuracy of 99.68%. The integrated Fuzzy C Means and EfficientNet model have been tested with various MRI images, and it outperforms the existing EfficientNet models in terms of accuracy.