An adaptive and altruistic PSO-based deep feature selection method for Pneumonia detection from Chest X-rays

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Pneumonia is a leading cause of child mortality, especially in low-income regions.
• Accurate pneumonia detection is challenging in developing countries due to limited resources.
• Computer-aided diagnosis (CAD) systems offer a cost-effective solution for medical diagnosis.

Purpose of the Study:

• To develop a novel CAD system for pneumonia detection from chest X-rays.
• To improve the accuracy and efficiency of pneumonia diagnosis using deep learning and meta-heuristic algorithms.
• To introduce an enhanced feature selection technique for better diagnostic performance.

Main Methods:

• Deep feature extraction using a fine-tuned ResNet50 model.
• Development of an adaptive and altruistic particle swarm optimization (AAPSO) for feature selection.
• Elimination of non-informative features to enhance diagnostic capabilities.

Main Results:

• The proposed AAPSO method significantly improved pneumonia detection accuracy.
• The system demonstrated superior performance compared to existing frameworks on a public pneumonia dataset.
• AAPSO showed effectiveness in diverse applications including cancer and COVID-19 prediction.

Conclusions:

• The developed CAD system, utilizing AAPSO for feature selection, offers a promising approach for accurate pneumonia detection.
• The method's efficiency and effectiveness are validated across various datasets, highlighting its potential for real-world applications.
• This research contributes to accessible healthcare solutions in resource-constrained environments.