Article: Ensemble of deep features and classifiers approach for MRI brain tumour classification Journal: International Journal of Intelligent Engineering Informatics (IJIEI) 2024 Vol.12 No.4 pp.433 - 459 Abstract: Medical professionals identify and classify brain tumours to save lives. This innovative study applies prominent machine learning classifiers to varied deep brain imaging features extracted by a pre-trained convolution neural network. Several machine learning classifiers use a pre-trained deep convolutional neural network's deep features to classify MRI images. Famous pre-trained networks extract MRI brain imaging properties. Multiple machine learning classifiers validate extracted traits. The finest deep features from numerous ML classifiers are assembled into feature sets and fed into multiple classifiers to predict classification. Pre-trained deep feature mining, machine learning classifiers, and brain tumour categorisation ensemble features are tested on BraTS-19, Figshare, and Kaggle datasets. Classifying brain tumour images as malignant or benign is difficult. To speed up categorisation, use ensemble deep features and a pre-trained model. Extraction of deep features from MRI images using transfer learning (EfficientNet-B4, Inception-V3, and VGG-19) is applied to popular classifiers (SVM, AdaBoost, Naïve Bayes, and random). The SVM radial basis function gets the top-3 traits from this method. This classification approach excels for huge MRI datasets. VGG-19 + Inception-V3 + EfficientNet-B4 on SVM-RBF classifier perform best on BraTS with 0.9673 accuracy. Inderscience Publishers - linking academia, business and industry through research

Title: Ensemble of deep features and classifiers approach for MRI brain tumour classification

Authors: B. Sathees Kumar

Addresses: Department of Computer Science, Bishop Heber College (Autonomous), Trichy, Tamil Nadu, India; Affiliated to Bharathidasan University, Trichy, Tamil Nadu, India

Abstract: Medical professionals identify and classify brain tumours to save lives. This innovative study applies prominent machine learning classifiers to varied deep brain imaging features extracted by a pre-trained convolution neural network. Several machine learning classifiers use a pre-trained deep convolutional neural network's deep features to classify MRI images. Famous pre-trained networks extract MRI brain imaging properties. Multiple machine learning classifiers validate extracted traits. The finest deep features from numerous ML classifiers are assembled into feature sets and fed into multiple classifiers to predict classification. Pre-trained deep feature mining, machine learning classifiers, and brain tumour categorisation ensemble features are tested on BraTS-19, Figshare, and Kaggle datasets. Classifying brain tumour images as malignant or benign is difficult. To speed up categorisation, use ensemble deep features and a pre-trained model. Extraction of deep features from MRI images using transfer learning (EfficientNet-B4, Inception-V3, and VGG-19) is applied to popular classifiers (SVM, AdaBoost, Naïve Bayes, and random). The SVM radial basis function gets the top-3 traits from this method. This classification approach excels for huge MRI datasets. VGG-19 + Inception-V3 + EfficientNet-B4 on SVM-RBF classifier perform best on BraTS with 0.9673 accuracy.

Keywords: deep learning; ensemble learning; transfer learning; machine learning; brain tumour classification; pre-trained deep convolutional neural network; recognition and categorisation.

DOI: 10.1504/IJIEI.2024.142417

International Journal of Intelligent Engineering Informatics, 2024 Vol.12 No.4, pp.433 - 459

Received: 27 Apr 2023
Accepted: 20 May 2024
Published online: 30 Oct 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Ensemble of deep features and classifiers approach for MRI brain tumour classification

Keep up-to-date