Tumor Detection and Analysis from Brain MRI Images Using Deep Learning

Abstract

Establishment of appropriate diagnosis plan for brain tumors is imperative and influences patient survival rate ultimately. Misdiagnosis of some brain tumors though hinders response to treatment and decreases success rate. Traditional ways of discerning brain tumors based on images taken via an MRI machine can be slow and full of mistakes, particularly when processing huge datasets and various shapes of tumor. It is proposed in this study that through such an innovative machine-learning and image processing hybrid technique, brain tumor detection can be enhanced. More precisely, we undertook bilateral filtering and normalization to improve quality of MR images, so as to decrease human errors in tumors’ identification. Therefore, the ResNet-50 model was employed to my significant patterns of data from the preprocessed dataset for further investigation. We did a comparative evaluation of different algorithms for getting the best classifier. Hence, based on the data, one can say with 91.40% accuracy that SVC outperformed all other models. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.