Adaptive Region of Interest (ROI) detection and tracking for respiration measurement in thermal video

Balvinder Kaur; Jill K. Nelson; Timothy Williams; Barbara L. O'Kane

doi:10.1117/12.922926

Translator Disclaimer

10 May 2012 Adaptive Region of Interest (ROI) detection and tracking for respiration measurement in thermal video

Balvinder Kaur, Jill K. Nelson, Timothy Williams, Barbara L. O'Kane

Author Affiliations +

Proceedings Volume 8401, Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering X; 840117 (2012) https://doi.org/10.1117/12.922926
Event: SPIE Defense, Security, and Sensing, 2012, Baltimore, Maryland, United States

Abstract

Respiration rate is a key guide for evaluating the physiological state of an individual during triage. Recent work has shown that high resolution thermal cameras can passively and remotely obtain respiration signals under controlled environmental conditions. This paper introduces an automatic end-to-end respiration signal measurement (through signal detection) approach based on statistical computation of the image intensities around the human nostril area in a thermal video. A method is presented to detect and track the nostril area and to calculate statistical values of the pixel intensity around the nostril area and correlate the statistical values with respiration signals from a contact sensor such as transducer belt. Results are based upon data collected from 200 subjects across two different experiments. This work provides not only a new image processing tool for tracking facial ROIs in thermal imagery, but also enhances our capability to provide non-contact, remote, passive, and real-time methods for measuring respiration for security and medical applications.

Citation Download Citation

Balvinder Kaur, Jill K. Nelson, Timothy Williams, and Barbara L. O'Kane "Adaptive Region of Interest (ROI) detection and tracking for respiration measurement in thermal video", Proc. SPIE 8401, Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering X, 840117 (10 May 2012); https://doi.org/10.1117/12.922926

ACCESS THE FULL ARTICLE