Augmenting Radiological Diagnostics with AI for Tuberculosis and COVID-19 Disease Detection: Deep Learning Detection...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer-Aided Diagnosis

Background:

• Chest X-rays are crucial for diagnosing lung conditions like pneumonia, COVID-19, and tuberculosis.
• Accurate and timely diagnosis is vital for effective patient treatment and disease management.
• Advanced deep learning models offer potential for improving the accuracy of image-based diagnoses.

Purpose of the Study:

• To develop and evaluate a deep learning network for the detection of pneumonia, COVID-19, and tuberculosis using chest X-ray images.
• To compare the performance of ResNet50 and VGG16 deep learning models in identifying these lung conditions.
• To assess the impact of data augmentation on model performance.

Main Methods:

• Utilized chest X-ray images for training and testing deep learning models, specifically VGG16 and ResNet50.
• Employed data augmentation techniques to enhance image datasets for improved model training.
• Evaluated model performance using precision, recall, F1 scores, and ROC AUC on training, validation, and testing datasets.

Main Results:

• The ResNet50 model significantly outperformed VGG16 in accuracy and resilience, showing superior ROC AUC values.
• ResNet50 achieved near 0.99 precision and recall for all conditions in the test set.
• VGG16 also demonstrated strong performance, with a precision of 0.99 and recall of 0.93 for tuberculosis detection.

Conclusions:

• Deep learning models, particularly ResNet50, show significant promise for accurate and reliable detection of lung conditions from X-ray images.
• The developed models can enhance diagnostic accuracy for diseases like COVID-19 and tuberculosis, potentially transforming clinical diagnostics.
• Data augmentation and model selection are key factors in advancing deep learning applications for medical imaging analysis.