Efficient Brain Tumor Detection with Lightweight End-to-End Deep Learning Model.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Science

Background:

• Deep learning, particularly Convolutional Neural Networks (CNNs), shows promise in medical imaging for brain tumor diagnosis.
• The Internet of Medical Things (IoMT) facilitates integrating deep learning into advanced diagnostic devices.
• Challenges include high computational costs and potential bias from insufficient training data in current deep learning models.

Purpose of the Study:

• To propose a novel, lightweight CNN-based deep learning model for brain tumor detection.
• To reduce system complexity and enable real-time applications in IoMT.
• To provide a framework for secure data transfer of medical results within the IoMT.

Main Methods:

• Developed a new, end-to-end, lightweight CNN model for brain tumor detection.
• Trained the model on medical imaging datasets for cancer recognition.
• Evaluated the model's performance for accuracy and efficiency in binary and multi-class scenarios.

Main Results:

• Achieved high accuracy rates: 99.48% for binary classification and 96.86% for multi-class classification.
• The proposed lightweight CNN model demonstrated superior performance compared to existing CNNs.
• The model's reduced complexity and small number of layers make it suitable for real-time applications.

Conclusions:

• The novel CNN model offers a significant advancement in deep learning for brain tumor detection within the IoMT.
• The model's efficiency and accuracy support its potential for real-time medical diagnostics.
• The study provides essential security recommendations for data transfer in IoMT environments.