Classification of Chronic Obstructive Pulmonary Disease (COPD) Through Respiratory Pattern Analysis.

Area of Science:

• Medical Imaging
• Respiratory Medicine
• Data Science

Background:

• Chronic Obstructive Pulmonary Disease (COPD) poses a significant global health challenge.
• Accurate and early diagnosis of COPD is crucial for effective patient management.
• Current diagnostic methods can be invasive or require specialized equipment.

Purpose of the Study:

• To propose a novel classification system for predicting COPD patients and non-patients.
• To leverage thermal imaging and respiratory pattern analysis for non-contact disease detection.
• To develop a data-driven approach for objective COPD assessment.

Main Methods:

• Respiratory volume was measured using thermal imaging.
• Key respiratory pattern features (volume, expiration/inspiration intervals, rate) were extracted and stored as text data.
• Features were normalized (Z-score) and combined via weighted summation to generate a classification score.
• Receiver Operating Characteristic (ROC) curve analysis determined the optimal threshold for patient classification.

Main Results:

• The classification system achieved an overall accuracy of 82.5%.
• Precision, recall, and F1-score metrics confirmed the model's high classification performance.
• The study successfully demonstrated the feasibility of using thermal imaging for COPD detection.

Conclusions:

• The proposed non-contact classification system shows significant potential for diagnosing respiratory diseases like COPD.
• This approach could facilitate remote medical examinations and improve accessibility to diagnostic tools.
• Further research may lead to widespread clinical application in non-invasive respiratory health monitoring.