Machine Learning based tumor diagnosis using compressive sensing in MRI images.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Signal Processing

Background:

• Magnetic Resonance Imaging (MRI) generates large datasets, posing challenges for timely disease diagnosis.
• Compressive Sensing (CS) offers a method to reconstruct images from fewer data points, potentially speeding up MRI analysis.
• Deep Learning (DL) models show promise in medical image analysis but often require extensive data.

Purpose of the Study:

• To investigate the efficacy of combining Compressive Sensing (CS) with Deep Learning (DL) for enhanced MRI analysis.
• To develop and evaluate a model for classifying MRI images as healthy or unhealthy using CS-reconstructed data.
• To assess the potential of CS-enhanced MRI for improving tumor diagnosis accuracy and efficiency.

Main Methods:

• A hybrid approach combining Compressive Sensing (CS) with the VGGNet-16 Deep Learning (DL) model was employed.
• MRI images, including normal and tumor datasets, were reconstructed using CS principles.
• The VGGNet-16 model was trained on CS-reconstructed images for binary classification (healthy vs. unhealthy).
• Performance was evaluated using metrics such as accuracy, precision, recall, and F1-score.

Main Results:

• The CS-enhanced VGGNet-16 model achieved a high classification accuracy of 98.7% for MRI images.
• The achieved accuracy is comparable to methods using traditionally acquired MRI data.
• The study demonstrated the feasibility of using CS for efficient MRI data acquisition and analysis.

Conclusions:

• Compressive Sensing (CS) combined with Deep Learning (DL) significantly improves the efficiency and accuracy of MRI-based tumor diagnosis.
• This approach holds potential for clinical applications, enabling faster and more reliable disease detection.
• Further research into CS applications across various medical imaging modalities is warranted to broaden diagnostic capabilities.