An efficient brain tumor detection and classification using pre-trained convolutional neural network models

Area of Science:

• Medical Imaging and Diagnostics
• Artificial Intelligence in Medicine
• Oncology

Background:

• Brain tumors are a significant cause of morbidity, necessitating accurate and timely diagnosis.
• Magnetic Resonance Imaging (MRI) is crucial for brain tumor detection but manual analysis is time-consuming and prone to errors.
• Deep Learning (DL) and Machine Learning (ML) offer potential for automated analysis of medical imaging data.

Purpose of the Study:

• To develop and evaluate a Convolutional Neural Network (CNN) model for accurate and efficient brain tumor detection from MRI scans.
• To improve upon manual assessment limitations by leveraging deep learning for faster and more precise cancer detection.
• To enhance diagnostic capabilities for radiologists, facilitating better patient treatment and decision-making.

Main Methods:

• Utilized a dataset comprising MRI scans classified into tumor types and non-tumor samples.
• Employed pre-trained CNN architectures, specifically ResNet50 and EfficientNet, for brain tumor classification.
• Implemented data augmentation techniques to increase dataset size and improve model robustness.

Main Results:

• The proposed CNN model, utilizing ResNet50 and EfficientNet, demonstrated high accuracy, precision, and recall in brain tumor detection.
• Evaluation metrics including validation loss, confusion matrix, and overall loss confirmed the model's effectiveness.
• The deep learning approach significantly reduced the time and potential inaccuracies associated with manual MRI assessment.

Conclusions:

• CNNs, particularly ResNet50 and EfficientNet, represent a powerful tool for automated brain tumor detection in MRI scans.
• This approach enhances diagnostic accuracy and efficiency, supporting radiologists in clinical decision-making.
• The study highlights the potential of deep learning to revolutionize brain cancer diagnosis and patient care.