MRI Brain tumor classification: application of various data augmentation techniques on ResNet50

One of the most fatal diseases, brain tumors have a significant impact on a patient's quality of life and survival rate. Effective therapy relies on early and precise detection, but standard diagnostic techniques, like MRI interpretation, are painstaking and prone to human error. Convolutional Neural Networks (CNNs), a type of deep learning approach, have shown impressive performance in medical picture classification; models such as ResNet50 have achieved high accuracy in image classification. This study investigates how data augmentation methods affect ResNet50's ability to classify brain tumors. To evaluate impact on classification performance, MRI dataset was tested using a variety of augmentation techniques. The findings show that augmentation methods increase model accuracy, with translation-based augmentations (79.95%) ranking in second and increasing sharpness providing the best accuracy (80.20%).