Brain tumor segmentation using deep neural image analysis
1 Lambton College, Toronto, Canada.
2 BE Computer Science, Purbanchal University, Nepal.
3 PIET, Kurukshetra University, India.
4 Lambton College, Toronto, Canada.
Research Article
World Journal of Advanced Research and Reviews, 2024, 24(03), 2873-2887
Article DOI: 10.30574/wjarr.2024.24.3.3970
Publication history:
Received on 16 November 2024; revised on 26 December 2024; accepted on 28 December 2024
Abstract:
This research investigates the application of deep learning techniques, specifically Convolutional Neural Networks (CNNs), for the classification of brain tumors in MRI images. The study aims to enhance diagnostic accuracy by leveraging the capabilities of CNNs to automatically learn spatial features from medical images, eliminating the need for manual feature extraction. The dataset used in this study includes MRI scans of brain tumors, where the model was trained and evaluated on the task of classifying tumors into different categories. The CNN architecture outperformed traditional machine learning methods and baseline models, such as VGG-16 and ResNet-50, achieving high accuracy, precision, recall, and F1-score, with a classification accuracy of 92.6%. Additionally, model interpretability was enhanced using Grad-CAM, which provided insights into the regions of interest in the MRI images, aiding in the model's decision-making process.
The study contributes to the growing body of knowledge in medical image analysis, demonstrating that deep learning models, particularly CNNs, can be an effective tool for brain tumor classification. The results highlight the model's potential for use in clinical settings, where accurate and rapid tumor detection is essential. However, the research also identifies limitations, including the need for larger and more diverse datasets and the challenge of overfitting. Future research directions include the exploration of 3D CNNs, multi-modal data fusion, and hybrid architectures to improve model performance. The study emphasizes the importance of continued efforts in enhancing model interpretability, integrating advanced AI techniques, and testing these models in real-world clinical environments to improve patient outcomes.
Keywords:
Brain Tumor; Segmentation; Deep Learning; Image Processing; CNN
Full text article in PDF:
Copyright information:
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0