Brain Tumor Classification: Leveraging Transfer Learning via EfficientNet-B0 Pretrained Model

Hegazy, Reem Tamer; Khalifa, Salmeen Khalil; Mortada, Rodina Amr; Amin, Basel A.; Elfattah, Amr Abd

doi:10.21608/iiis.2025.292441.1034

Brain Tumor Classification: Leveraging Transfer Learning via EfficientNet-B0 Pretrained Model

Document Type : Original Article

Authors

¹ Department of Artificial Intelligence, Misr university for science and technology

² Departament of computer science,Misr university for science and technology

³ Department of artificial intelligence,Misr university for science and technology

⁴ Department of computer science, Misr university for science and technology

10.21608/iiis.2025.292441.1034

Abstract

Abstract— Brain tumor classification from MRI scans is an essential task in medical diagnostics, enhancing the precision and speed of treatment planning. This project introduces a deep learning model that automates the classification of brain tumors by leveraging a pre-trained convolutional neural network (CNN). The model processes MRI images and categorizes them into one of four possible classes: glioma, meningioma, pituitary tumor, or no tumor. By utilizing the EffnetB0 pretrained model, our approach benefits from learned features on a broad range of visual data, allowing for robust feature extraction even with a limited number of medical images. The dataset consists of MRI scans, each labeled according to the tumor type(glimona, meningioma, no tumor,pitutary ), facilitating supervised learning. The effectiveness of the model is assessed based on accuracy, precision, and recall metrics, aiming to support radiologists by providing a reliable preliminary diagnostic tool that improves the diagnostic workflow for brain tumors.

Keywords