Efficient Deep Learning Approach for Detection of Brain Tumor Disease

Authors:
Open Access
Publisher website
Google Scholar
Abstract
This article presents, we describe fully autonomous brain tumor classifying as well as the classification method depending on a multiscale DCNN Deep-Convolutional-Neural-Network (multiscale Deep Convolutional Neural Network). As compared to similar efforts, ours differs in part because it processes input images at two different spatial scales using independent neural networks. The suggested neural model can interpret MRI scans comprising two kinds of tumors: the first one meningioma, and the second one glioma, without requiring input images to be preprocessed to eliminate skull and spine components. The effectiveness we've noted model available to the public MRI image dataset containing 3060 slices of 250 cases is compared to previously reported models for conventional ML (machine learning) as well as DL (deep learning). In the comparison, our approach produced an impressive 97% accuracy in classifying tumors, which was better than some other algorithms utilizing similar datasets.

This publication has 25 references indexed in Scilit: