A Comparative Study of DL and ML Models for Brain Tumor Detection

Brain tumor detection is a critical component of modern health care, with early and accurate diagnosis significantly impacting patient outcomes. Brain tumors are often detected and diagnosed using imaging techniques including CT scans, radiography, and MRI. In this study, we investigate the effectiveness of various state-of-the-art deep learning models, including VGG16, AlexNet, and CNN as well as machine learning models, including RF, SVM, and KNN in the perspective of brain tumor detection. Datasets used for the study, namely Brain Tumor Image Segmentation Benchmark and glioma, are exploited to test them with respect to major parameters: accuracy, precision, and recall. VGG16 and AlexNet are effective in capturing complex image attributes and thus considered in this work for analyzing and classifying brain tumor images, and the Convolutional Neural Network algorithm is employed as a reference. The comparative analysis quantifies and exhibits the efficiency and limitations of these algorithms. VGG16 achieved remarkable results with a precision of 93.27%, recall of 93.72%, accuracy of 95.23%, and an F1-score of 94.22%.