Level-Set Segmentation-Based Respiratory Volume Estimation Using a Depth Camera.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Respiratory Physiology

Background:

• Accurate measurement of human respiratory volume is crucial for clinical monitoring and respiratory support.
• Existing methods for respiratory volume measurement may have limitations in accuracy or non-invasiveness.

Purpose of the Study:

• To propose and validate a novel method for measuring human respiratory volume using a depth camera.
• To assess the accuracy and reliability of the proposed method compared to existing level-set approaches.

Main Methods:

• Utilized a depth camera to capture chest wall movement.
• Employed a level-set segmentation technique incorporating spatial (chest wall shape) and temporal (previous frame segmentation) information.
• Validated the method against a ventilator and compared it with other level-set variations.

Main Results:

• The proposed method achieved a mean tidal volume error of 8.41%, significantly lower than spatial-only (14.34%) and temporal-only (10.93%) level-set methods.
• Demonstrated a high average intra-class correlation coefficient (ICC) of 0.893 for respiratory volume waveform compared to a ventilator.
• Achieved statistically significant improvements in tidal volume error reduction.

Conclusions:

• The proposed depth camera-based level-set method offers a highly accurate and reliable approach for measuring human respiratory volume.
• This non-invasive technique shows potential for improved respiratory monitoring in clinical settings.
• The integration of spatial and temporal information enhances the precision of respiratory volume quantification.