Flow-Field Inference for Turbulent Exhale Flow Measurement

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Pulmonary Diagnostics

Background:

• Remote imaging offers a unique approach to pulmonary diagnostics by tracking natural breathing behaviors.
• Existing methods often rely on chest and diaphragm movements, but this study focuses on thermal carbon dioxide (CO2) visualization of exhale flow.

Purpose of the Study:

• To develop a novel respiratory diagnostic tool for capturing and quantifying natural breathing using thermal CO2 exhale flow patterns.
• To identify and measure key respiratory metrics including breathing rate, flow, and volume.
• To establish a method for capturing individualized pulmonary traits through subtle respiratory signature analysis.

Main Methods:

• A novel method was developed to isolate and extract turbulent exhale flow signals from thermal image sequences.
• Flow-field prediction and optical flow measurement techniques were employed.
• An inference model (FieldNet) was trained using clinical data to quantify exhale behaviors.

Main Results:

• Expiratory flow measurements revealed individualized flow signatures in the initial cohort.
• The proposed flow field model successfully isolated and analyzed turbulent exhale behaviors.
• The system demonstrated the ability to measure anomalous respiratory behaviors.

Conclusions:

• Detailed spatial flow analysis of thermal CO2 patterns can identify unique, patient-specific breathing signatures and abnormalities.
• This research represents a significant step towards non-contact respiratory monitoring.
• The technology directly measures effort-independent behaviors by analyzing exhaled CO2 airflow.