Optimal and Efficient Deep Learning Model for Brain Tumor Magnetic Resonance Imaging Classification and Analysis
Open Access
- 8 August 2022
- journal article
- research article
- Published by MDPI AG in Applied Sciences
- Vol. 12 (15) , 7953
Abstract
A brain tumor (BT) is an abnormal development of brain cells that causes damage to the nerves and blood vessels. An accurate and early diagnosis of BT is important to prevent future complications. Precise segmentation of the BT provides a basis for surgical and planning treatment to physicians. Manual detection utilizing MRI images is computationally difficult. Due to significant variation in their structure and location, viz., ambiguous boundaries and irregular shapes, computerized tumor diagnosis is still a challenging task. The application of a convolutional neural network (CNN) helps radiotherapists categorize the types of BT from magnetic resonance images (MRI). This study designs an evolutional algorithm with a deep learning-driven brain tumor MRI image classification (EADL-BTMIC) model. The presented EADL-BTMIC model aims to accurately recognize and categorize MRI images to identify BT. The EADL-BTMIC model primarily applies bilateral filtering (BF) based noise removal and skull stripping as a pre-processing stage. In addition, the morphological segmentation process is carried out to determine the affected regions in the image. Moreover, sooty tern optimization (STO) with the Xception model is exploited for feature extraction. Furthermore, the attention-based long short-term memory (ALSTM) technique is exploited for the classification of BT into distinct classes. To portray the increased performance of the EADL-BTMIC model, a series of simulations were carried out on the benchmark dataset. The experimental outcomes highlighted the enhancements of the EADL-BTMIC model over recent models.Keywords
Funding Information
- King Khalid University (25/43, PNURSP2022R114, 22UQU4210118DSR29)
- Princess Nourah Bint Abdulrahman University (25/43, PNURSP2022R114, 22UQU4210118DSR29)
- Umm Al-Qura University (25/43, PNURSP2022R114, 22UQU4210118DSR29)
This publication has 21 references indexed in Scilit:
- Segmentation and classification of brain tumors using modified median noise filter and deep learning approachesMultimedia Tools and Applications, 2021
- Classification of brain tumors from MR images using deep transfer learningThe Journal of Supercomputing, 2021
- Joint training of two-channel deep neural network for brain tumor classificationSignal, Image and Video Processing, 2020
- Brain Tumor Identification and Classification of MRI Images Using Deep Learning TechniquesIEEE Access, 2020
- A Novel Hybrid Deep Learning Model for Sentiment ClassificationIEEE Access, 2020
- Brain tumor classification in MRI image using convolutional neural networkMathematical Biosciences and Engineering, 2020
- A Deep Learning-Based Framework for Automatic Brain Tumors Classification Using Transfer LearningCircuits, Systems, and Signal Processing, 2019
- Brain tumor classification for MR images using transfer learning and fine-tuningComputerized Medical Imaging and Graphics, 2019
- A Review on a Deep Learning Perspective in Brain Cancer ClassificationCancers, 2019
- Classification using deep learning neural networks for brain tumorsFuture Computing and Informatics Journal, 2018