Brain Tumor Detection Using Deep Learning Algorithms

Abstract

Brain tumors are a serious health concern and also rank among the prime causes of cancer deaths. Diagnosis has to be done as early as possible and with accuracy so that treatment may start at an early stage for better survival rates. MRI happens to be the modality of choice in imaging brain tumors because of its wide application in the diagnosis of this condition. However, the detection of such diseases through conventional methods, including machine learning algorithms, Fuzzy C-means, and Artificial Neural Networks, usually has serious drawbacks regarding reduced accuracy and computational time. This work will propose a new detection scheme based on a sequential deep learning algorithm combined with Long Short-Term Memory networks that will improve diagnostic precision. The LSTM model tended to be the best, often reaching accuracy as high as 89%, hence being the potential promise for brain tumor detection from MRI images more quickly and with higher reliability.