Running summarizing tools on a new article

Check the paper below, then start its synthesis.


Brain Tumor Detection using Transfer Learning

Mehek Jagwani (Delhi Technological University), Shweta Meena (Delhi Technological University)

International Conference on Computational Engineering, Sensing Technology and Management (ICCETM)
License: CC BY

Abstract: Brain tumors give rise to a serious global public health problem, necessitating timely and accurate detection to allow for effectual treatment and improved patient outcome. Magnetic Resonance Imaging (MRI) scan analysis by humans is often time-consuming, subject to human error, and also, request technical expertise. This research aims to address these restraints through the development of high-performance automatic brain tumor classification system using transfer learning under the direction of VGG16 deep learning model. Our perspective lever-ages on the pre-trained VGG16 model, which was fine-tuned on a huge MRI image dataset, to classify tumors into four groups as: glioma, meningioma, pituitary, and no tumor. Data augmentation approaches were highly used during the training phase as an attempt to uplift the overall generalization and stability of the model. The suggested model attain a tremendously high test accuracy of 96%, with outstanding dimensions of discrimination among the various tumor groups and normal brain tissues, with exceptionally powerful performance in the detection of ’no tumor’ group. Inspite of the fact that other modern deep architectures like ResNet and DenseNet carry out satisfactory presentation in this direction, our contribution is a productive and well organized single-model transfer learning system. This contribution come up with a good base to authorize clinical decision-making, improve diagnosis set backs, and act as a foundation for future attempts like architectural alteration or ensemble methods to additionally intensify the performance.

Published on 18 Nov. 2025