A Deep Learning Approach for Brain Tumor Classification and Segmentation Using a Multiscale Convolutional Neural Network

Abstract

In this paper, we present a fully automatic brain tumor segmentation and classification model using a Deep Convolutional Neural Network that includes a multiscale approach. One of the differences of our proposal with respect to previous works is that input images are processed in three spatial scales along different processing pathways. This mechanism is inspired in the inherent operation of the Human Visual System. The proposed neural model can analyze MRI images containing three types of tumors: meningioma, glioma, and pituitary tumor, over sagittal, coronal, and axial views and does not need preprocessing of input images to remove skull or vertebral column parts in advance. The performance of our method on a publicly available MRI image dataset of 3064 slices from 233 patients is compared with previously classical machine learning and deep learning published methods. In the comparison, our method remarkably obtained a tumor classification accuracy of 0.973, higher than the other approaches using the same database.

A Novel Deep Learning Method for Recognition and Classification of Brain Tumors from MRI Images

“…In this section, we compare our proposed model with the other segmentation techniques using the Figshare brain tumor dataset [ 38 ] which is one of the largest online available datasets. For performance evaluation, we compared the average highest results of our presented technique with the average results reported in these studies [ 61 , 62 , 63 , 64 , 65 ]. For the presented technique, we have shown the results for the DenseNet-41-based Mask-RCNN framework as we obtained better performance on it as compared to the ResNet-50 framework.…”

“…However, the authors considered only the axial MR images of glioma and meningioma brain tumors. In [ 64 ], the authors proposed a multi-scale CNN model that processes the input MR image in three different spatial scales using multiple processing pathways. They achieved an average accuracy of 0.973 and a dice score of 0.828 for classification and segmentation, respectively.…”

“…The proposed technique uses deep features that are more discriminating, reliable, and provide a more effective representation of tumor regions over other methods such as [ 61 ], which employs the hand-crafted features and is unable to better represent the tumor region due to structural complexities. Moreover, in some existing methods [ 62 , 64 ] segmentation is applied directly to the entire image, which results in misclassification due to the complex background (i.e., brain tissues overlapping with tumor boundary, MRI artifacts, etc. ), which thus reduces the accuracy of the segmentation.…”

Computer-Aided Brain Tumor Diagnosis: Performance Evaluation of Deep Learner CNN Using Augmented Brain MRI

“…Using 5-fold cross-validation, average accuracy achieved by the proposed technique is 91.46%. A multiscale Deep CNN [ 29 ] is proposed which can analyse tumor MRIs and classify them into glioma, meningioma, and pituitary tumor. The performance of the proposed model is evaluated on an MRI image dataset consisting of 3,064 images.…”

Development of a Fully Automated Glioma-Grading Pipeline Using Post-Contrast T1-Weighted Images Combined with Cloud-Based 3D Convolutional Neural Network

“…2021, 11, 5118 2 of 11 CNN is being developed. Abd El Kader et al [8] proposed a differential deep CNN model to classify abnormal or normal MR brain images, and Díaz-Pernas et al [9] proposed a multiscale CNN model for region extraction and classification of three types of brain tumors, including glioma, in post-contrast T1-weighted images. However, in these studies, 3D MR images were analyzed on a slice-by-slice basis.…”