Long term, non-contact monitoring allows for continuous and consistent health measurements. The use of multi-modality systems to monitor vital signals of a given subject, should result in less error and require less processing time. Context-awareness has the potential to facilitate the use of several modalities in one system. This paper examines the viability of context-awareness in a multi-modality system to facilitate system synergy. This is achieved by extracting pulse measurements from both modalities and examining the extracted signals as one set of algorithms is forced to fail. Thermal and visible light video data was collected from one subject as they sat quietly. The visible light video was augmented to present with diminishing light over time. The thermal and augmented visible light video data was then subjected to pre-processing, data segmentation and Adaptive Eulerian Video Magnification before extracting pulse measurements. The data processing methods were successful in identifying and enhancing signals representative of pulse. The visible-light algorithms failed to extract pulse measurements when the light became too dim, while the thermal algorithms continued to extract pulse measurements. These results suggest that context-awareness may improve error and processing time by detecting system failures, and accounting for them. In examining signal accuracy, the thermal and visible light pulse signals generally correlated well with the gold standard, Respiratory Inductance Plethysmography (RIP) data, with some differences in phase. These results suggest that the algorithm can be optimized using techniques such as region tracking. Future work involves system optimization and multi-modality signal fusion.

Additional Metadata
Keywords Context aware, heart rate, non-contact, thermal imaging
Persistent URL dx.doi.org/10.1109/MeMeA.2018.8438684
Conference 13th IEEE International Symposium on Medical Measurements and Applications, MeMeA 2018
Citation
Bennett, S.L. (Stephanie L.), Goubran, R, & Knoefel, F. (2018). Context-Awareness in Non-Contact, Multi-Modality, Bed-Based Monitoring of Vital Signs. In MeMeA 2018 - 2018 IEEE International Symposium on Medical Measurements and Applications, Proceedings. doi:10.1109/MeMeA.2018.8438684