Developing an Optimize Deep Learning Framework for Brain Tumors Detection and Classification

¹ Research Scholar, Institute of Engineering and Management, University of Engineering and Management, Kolkata, India
² Research Scholar, Swami Vivekananda University, Kolkata, India
³ Associate Professor, MBA Dept., Institute of Engineering and Management, University of Engineering and Management, Kolkata, India
⁴ Final year MBA Student, Institute of Engineering and Management, Kolkata, India

^* Corresponding author: prasenjit.kundu@iem.edu.in

Published online: 5 May 2025

Abstract

Brain tumors are considered one of the present health challenges. Perhaps the early spotting could be an important factor in their effective treatment and in the good outcome of the patients. But, due to the asymptomatic nature and location of brain tumor, early detection through tissue diagnosis is difficult. This research design an empirical deep learningbased framework to detect and classify brain tumors from Magnetic Resonance Imaging (MRI) images datasets. The research work aims at improving the quality and reliability of radiological images through preprocessing techniques such as resizing, noise reduction, normalization, and segmentation to classify tumors more precisely. The aim of this study to develop hybrid deep learning based model to ensure that benign and malignant brain tumors are appropriately differentiated and also attempt to provide a framework for predicting the type of malignant tumors. This article evaluated three deep learning frameworks on radiological tumor images dataset to train the model and based on accuracy, validation and performance, the most optimize model is selected to apply on test dataset to automatically classified brain tumors. Authors of our study are hopeful for adding value to diagnostic precision. Automation in the classification process optimizes efficiency for faster and more reliable diagnostic support. Future research are expected to incorporate multi-modal data to enhance the diagnostic accuracy and smoothness of its clinical application. This may most likely improve the speed and precision of the diagnosis made by health professionals using deep learning (DL) in the diagnosis of brain tumors and help to improve patient care.