Klasifikasi Tumor Otak Berbasis Data Mri dengan Algoritma Convolutional Neural Network (CNN)

Abstract

Brain tumors require accurate and timely diagnosis to support appropriate medical treatment. This study proposes a brain tumor classification system based on Magnetic Resonance Imaging (MRI) using a Convolutional Neural Network (CNN). The dataset consists of 5,266 MRI images categorized into three classes: glioma, meningioma, and no tumor. Prior to classification, image preprocessing was performed through image resizing, normalization, and Local Binary Pattern (LBP) feature extraction. The dataset was divided into 3,403 training images, 852 validation images, and 1,011 testing images. The proposed CNN architecture consists of three convolutional layers, max pooling layers, global average pooling, a dense layer, dropout, and a softmax output layer. Experimental results show that the model achieved training, validation, and testing accuracies of 98.53%, 94.24%, and 92.28%, respectively. These results indicate that CNN can effectively classify brain MRI images and provide good generalization performance for unseen data.