Deep-Pneumonia Framework Using Deep Learning Models Based on Chest X-Ray Images.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Science

Background:

• Pneumonia is a leading cause of death globally, particularly in vulnerable populations like children and the elderly.
• Accurate and efficient detection of pneumonia from chest X-rays is crucial for timely treatment.
• Existing deep learning models face challenges in meeting all performance metrics for pneumonia detection.

Purpose of the Study:

• To develop and evaluate efficient deep learning models for detecting and classifying pneumonia from chest X-ray images.
• To compare the performance of different deep learning architectures, including pre-trained and custom models.

Main Methods:

• Four deep learning models were developed: ResNet152V2, MobileNetV2, a Convolutional Neural Network (CNN), and a Long Short-Term Memory (LSTM) integrated with CNN.
• Models were implemented and evaluated using Python.
• Performance was compared against recent research in the field.

Main Results:

• The ResNet152V2 model achieved exceptional performance metrics: 99.22% accuracy, 99.43% precision, 99.44% F1-score, 99.44% recall, and 99.77% Area Under the Curve (AUC).
• MobileNetV2, CNN, and LSTM-CNN models also demonstrated strong results, exceeding 91% in accuracy, recall, F1-score, precision, and AUC.
• The proposed deep learning framework, particularly ResNet152V2, outperformed recently published models.

Conclusions:

• Deep learning models, especially ResNet152V2, offer a powerful and accurate approach for pneumonia detection from chest X-rays.
• The developed models provide significant improvements in key performance metrics, potentially aiding clinical diagnosis.
• The study highlights the effectiveness of advanced deep learning architectures in medical image analysis for pneumonia.