Impact of Data Augmentation Technique on Automatic Brain Tumor Detection Using ResNet CNN

The research article studies the impact of data augmentation, particularly 60˚ rotation, on the performance of a ResNet convolutional neural network (CNN) for brain tumor detection using MRI images. A limited Kaggle dataset was used, supplemented with augmented images. The study compares model performance with and without augmentation, analyzing metrics such as accuracy, precision, recall, F1 score, and observing training and validation curves. The results indicate that data augmentation significantly improved performance, increasing accuracy from 63% to 90%.