Brain Image Segmentation for Tumor Detection Using Ensemble Machine Learning Techniques

Brain tumors can develop when abnormal cells in the brain grow uncontrollably, and MRI images provide valuable information about the presence of unwanted tissue growth. Numerous academic publications have utilized machine learning and deep learning algorithms to detect malignant tumors. These algorithms can swiftly identify brain tumors in MRI scans, leading to faster prediction times and improved precision, thereby facilitating easier patient treatment. Radiologists benefit from the accelerated decision-making process enabled by these predictive models. The proposed study evaluates the performance of Rectified Linear Unit (ReLU) and Sigmoid activation function-based Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN) networks for the detection of brain tumors.

I. INTRODUCTION

Recent advancements in technology have allowed for detailed examination of the human brain through MRI, which can be used for various types of research. Medical picture segmentation involves identifying and naming anatomical components of the brain, but requires manual segmentation masks and domain specialists for annotations. Performance may be subpar when distinguishing between malignant and benign tissues [1] [2] [14].