Frequent CNN based ensembling for MRI classification for Abnormal Brain Growth detection

Abstract

Digital image processing is a key player in the analysis of medical images, particularly in understanding the intricacies of abnormal brain growth development. Notably, the application of CNN algorithms to MRI images accelerates abnormal brain growth detection with enhanced accuracy; facilitating prompt decision-making by radiologists. This research focuses on finding abnormal brain growth using advanced CNN computer techniques. The study is split into three main steps. In the first step, brain MRI images are pre-processed by applying selected pre-processing techniques. In the second step, machine learning feature extraction methods are applied to pick out important features from these images. Finally, CNN models such as VGG, ResNet, DenseNet, and MobileNet are applied to classify the MRI images at a detailed level. The ensemble is done to improve the accuracy of the classification of MRI images. The results from study indicate easy automated abnormal brain growth detection that save radiologists' time and improve the efficiency of early diagnosis.

URN:NBN:sciencein.jist.2024.v12.785