Application of Machine Learning Techniques for the Location and Early Detection of Brain Tumors

This research introduces a deep learning method for identifying, classifying, and localizing brain tumors through MRI images. A Convolutional Neural Network (CNN) utilizing transfer learning is used to improve classification accuracy while tackling data constraints. In particular, the Inception model is utilized for classification, whereas NasNet-Mobile is applied for accurate tumor localization. The study employs a publicly accessible MRI dataset that includes images of meningiomas, gliomas, pituitary tumors, and non-tumorous cases. Techniques for preprocessing, including image normalization, augmentation, and annotation, were utilized to enhance the model's robustness. The data was divided into 80 % for training and 20 % for evaluation. The classification model reached a training accuracy of 99 % and a test accuracy of 78 %. The localization model showed excellent accuracy, achieving a mean square error (MSE) of 0.01 in training and 0.03 in the testing dataset.