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

Abstract

In this paper, a method is proposed by which to measure human respiratory volume using a depth camera. The level-set segmentation method, combined with spatial and temporal information, was used to measure respiratory volume accurately. The shape of the human chest wall was used as spatial information. As temporal information, the segmentation result from the previous frame in the time-aligned depth image was used. The results of the proposed method were verified using a ventilator. The proposed method was also compared with other level-set methods. The result showed that the mean tidal volume error of the proposed method was 8.41% compared to the actual tidal volume. This was calculated to be less error than with two other methods : the level-set method with spatial information (14.34%) and the level-set method with temporal information (10.93%). The difference between these methods of tidal volume error was statistically significant ($p \lt 0.0001$). The intra-class correlation coefficient (ICC) of the respiratory volume waveform measured by a ventilator and by the proposed method was 0.893 on average, while the ICC between the ventilator and the other methods were 0.837 and 0.879 on average.

Contact-free assessments of respiratory rate and volume with load cells under the bed legs in ventilated patients: a prospective exploratory observational study

“…There are growing interests and efforts in development of contact-free respiratory monitors, particularly after COVID-19; however, most of these studies tested only respiratory rates and did not assess tidal volumes. Furthermore, they were performed in nonclinical situations and for very limited short periods, under supervision of the researchers, when the best performance of the devices was possible ( 14 – 18 ). Depth camera and radar have been reported to produce excellent performance in measuring tidal volume without subject contact of human volunteers; however, clinical implementation of these technologies appears to be difficult because of difficulty in maintaining different body positions of patients covered by blankets ( 14 – 17 ).…”

“…Furthermore, they were performed in nonclinical situations and for very limited short periods, under supervision of the researchers, when the best performance of the devices was possible ( 14 – 18 ). Depth camera and radar have been reported to produce excellent performance in measuring tidal volume without subject contact of human volunteers; however, clinical implementation of these technologies appears to be difficult because of difficulty in maintaining different body positions of patients covered by blankets ( 14 – 17 ). To our knowledge, only one study has systematically assessed performance of tidal volume measurements with load cells under the bed legs ( 18 ).…”

Noncontact Respiratory Monitoring Using Depth Sensing Cameras: A Review of Current Literature

“…Although no statistical measures of error were provided in their paper, the per-subject scatter and corresponding Bland–Altman plots with limits of agreement drawn on are provided. Oh et al [ 35 ] studied 10 healthy adult volunteers, comparing their results against a ventilator reference. They obtained a correlation coefficient and mean tidal volume error of 0.98 and 8.1%, respectively, when they combined both spatial and temporal information within their method.…”

Continuous non‐contact respiratory rate and tidal volume monitoring using a Depth Sensing Camera

“…They do so by first deriving a respiratory volume (RV) signal from the respiratory motions of the patient from which these parameters can be extracted. The non-contact monitoring of RR and TV would prove valuable in the monitoring of viral pandemics, including novel coronavirus (COVID- 19) patients, as well as those with other viral respiratory tract diseases, where minimum contact with the patient is desired and a robust measurement is essential [6,7].…”

“…(Per-subject scatter plots and corresponding Bland-Altman plots with limits of agreement drawn on were provided in the paper, although no statistical measures of error such as mean bias, RMSD, or limits of agreement were stated by the authors.) Oh et al [19] studied 10 healthy adult volunteers comparing the depth tidal volume against a ventilator reference. They obtained a correlation coefficient and mean tidal volume error of 0.98 and 8.1% respectively.…”