eng Oriental Scientific Publishing Company Biomedical and Pharmacology Journal 0974-6242 2025-10-21 18 October Spl Edition 68496 article Mohammed Muddasir Naseer 1 Veeraprathap Veerabhadraiah 2 Basavaraj Rayappa Ramji 3 Santhosh Kumar Kadur Lokeshappa 4 Srividya Chandagirikoppal Nagendra 5 Yathiraj Guduganahalli Ramesh 6 Department of ISE, Vidyavardhaka College of Engineering, Mysuru, Karnataka, India Department of ECE, ATME College of Engineering, Mysuru, Karnataka, India Department of Industrial Engineering and Management, BMS College of Engineering, Bangalore, Karnataka, India School of Computer Science and Engineering, Presidency University, Bengaluru, Karnataka, India Department of ECE, BGS Institute of Technology, Adichunchanagiri University, Karnataka, India Department of CSE (Cyber Security), Coorg Institute of Technology, Ponnampet, Karnataka, India Magnetic resonance imaging (MRI) is widely used for the non-invasive diagnosis of brain tumours; however, accurate manual interpretation remains challenging due to the complexity of tumour structures and inter-observer variability. In this study, we propose an optimized EfficientNet-B0 architecture for automated brain tumour detection and classification. Two models were developed: Model 1, trained from scratch, and Model 2, which integrated transfer learning, fine-tuning, and hyperparameter optimization. The models were trained and evaluated on a publicly available brain MRI dataset comprising four categories: glioma, meningioma, pituitary tumour, and no tumour. Model 1 achieved a testing accuracy of 94.52%, while Model 2 outperformed it with 99.43% accuracy, demonstrating superior generalization and stability. Comparative analysis with previous studies further confirmed that the proposed approach achieves higher accuracy and robustness in multi-class classification tasks. The results indicate that the optimized EfficientNet-B0 with transfer learning provides a reliable and effective framework for clinical decision support in brain tumour diagnosis. https://biomedpharmajournal.org/vol18octoberspledition/an-optimized-efficientnet-b0-framework-for-multi-class-brain-tumour-detection-and-classification-from-mri-images/ Accuracy Brain Tumour Classification Convolutional Neural Network (CNN) Deep Learning EfficientNet-B0 MRI Image Analysis