Non-contact physiological monitoring of preterm infants in the Neonatal Intensive Care Unit

Abstract

The implementation of video-based non-contact technologies to monitor the vital signs of preterm infants in the hospital presents several challenges, such as the detection of the presence or the absence of a patient in the video frame, robustness to changes in lighting conditions, automated identification of suitable time periods and regions of interest from which vital signs can be estimated. We carried out a clinical study to evaluate the accuracy and the proportion of time that heart rate and respiratory rate can be estimated from preterm infants using only a video camera in a clinical environment, without interfering with regular patient care. A total of 426.6 h of video and reference vital signs were recorded for 90 sessions from 30 preterm infants in the Neonatal Intensive Care Unit (NICU) of the John Radcliffe Hospital in Oxford. Each preterm infant was recorded under regular ambient light during daytime for up to four consecutive days. We developed multi-task deep learning algorithms to automatically segment skin areas and to estimate vital signs only when the infant was present in the field of view of the video camera and no clinical interventions were undertaken. We propose signal quality assessment algorithms for both heart rate and respiratory rate to discriminate between clinically acceptable and noisy signals. The mean absolute error between the reference and camera-derived heart rates was 2.3 beats/min for over 76% of the time for which the reference and camera data were valid. The mean absolute error between the reference and camera-derived respiratory rate was 3.5 breaths/min for over 82% of the time. Accurate estimates of heart rate and respiratory rate could be derived for at least 90% of the time, if gaps of up to 30 seconds with no estimates were allowed.

Non-contact vital-sign monitoring of patients undergoing haemodialysis treatment

“…Figure 4 shows the distribution of the time gaps between heart rate estimates from the video camera for the entire dialysis data set. The majority of gaps between the estimates were under 3 min, with the most common gap lasting for 30 s. Although some longer gaps occurred occasionally in the dialysis population, our results are comparable to our previous clinical studies involving the monitoring of vital signs of preterm infants in the Neonatal Intensive Care Unit (NICU) 14 . We reported that the majority of time periods for which heart rate could not be computed in the NICU were less than 30 s.

Figure 4 Gaps in heart rate estimation using the Kalman ND data fusion method.

…”

“…Some factors to consider are the quality of the lighting environment available in the hospital ward, the field of view required, the distance at which the camera would be set and the number of patients to be simultaneously monitored from a single camera. If only one patient is required to be monitored at a close distance in a well-lit scenario, such as a camera on top of a neonatal incubator to monitor infants in the NICU 14 , a low-resolution off-the-shelf camera can potentially be used. However, the further the camera is placed from the patient(s), the higher the requirements of the quality of the camera will be.…”

Reference Free Incremental Deep Learning Model Applied for Camera-Based Respiration Monitoring

“…methods (PCA and ICA blind separation techniques) have been disclosed in [14] based on a similar database having MAE 6.9 BMP and 7.5 BPM values. The [16] describes an implementation of video-based non-contact technologiy to monitor the vital signs of preterm infants. The used dataset is composed of several patients and long recording time.…”

“…The involved network was trained by manually annotated skin regions and reached mean absolute error (MAE) 6.9-7.5 breaths per seconds in the collected data set including motion active regions as well. The [15] and [16] demonstrate a complex monitoring system based on a multiple output convolutional neural network. The solution gives a higher level detection capability such as intervention periods.…”

Combining Chrominance Features and Fast ICA for Noncontact Imaging Photoplethysmography

“…Recent studies have shown that heart rate can be determined by analyzing skin color changes in video signals. This noncontact and inexpensive method can also be used in certain special scenarios, such as skin damage, neonatal care, and imperceptible monitoring situations [6]- [8]. With the continuous development of noncontact measurement technology, imaging photoplethysmography (PPG)iPPG can provide convenience and comfort and reduce medical costs given that it only requires the processing of body skin color from videos to monitor heart rates automatically by means of data fusion and analysis.…”