AI Based Diagnosis of Pneumonia

Area of Science:

• Medical Diagnostics
• Artificial Intelligence in Healthcare
• Respiratory Medicine

Background:

• Pneumonia is a common lung infection causing inflammation of the alveoli, leading to breathing difficulties.
• Current diagnostic methods include chest X-ray, blood tests, and pulse oximetry.
• Lung sounds recorded via digital stethoscopes offer potential for non-invasive pneumonia detection.

Purpose of the Study:

• To develop and evaluate a software tool for diagnosing pneumonia using lung sound analysis.
• To assess the efficacy of the gradient boosting algorithm in identifying pneumonia from audio data.
• To compare the performance of the proposed method against other machine learning and deep learning techniques.

Main Methods:

• Lung sounds were recorded using electronic stethoscopes and processed with Audacity software to remove noise.
• A dataset of healthy and pneumonia patient lung sound recordings was created and labeled.
• The gradient boosting algorithm was trained on the labeled data for pneumonia classification.

Main Results:

• The gradient boosting algorithm achieved 97% accuracy in diagnosing pneumonia from lung sounds.
• The developed software demonstrated superior performance compared to support vector machine and k-nearest neighbors algorithms.
• The method showed excellent diagnostic capabilities, outperforming other artificial intelligence and deep learning approaches.

Conclusions:

• Gradient boosting algorithm effectively diagnoses pneumonia from lung sounds with high accuracy.
• This AI-powered software provides a promising, non-invasive tool for pneumonia detection.
• The method holds potential for predicting lung sound alterations associated with COVID-19.