Computer Vision Approach for Estimating Human Health Parameters

Measurement of vital cardiovascular health attributes, e.g., pulse rate variability, and estimation of exertion level of a person can help in diagnosing potential cardiovascular diseases, musculoskeletal injuries and thus monitoring an individual's well-being. Cumulative exposure to repetitive and forceful activities may lead to musculoskeletal injuries which not only reduce workers' efficiency and productivity, but also affect their quality of life. Existing techniques for such measurements pose a great challenge as they are either intrusive, interfere with human-machine interface, and/or subjective in the nature, thus are not scalable. Non-contact methods to measure these metrics can eliminate the need for specialized piece of equipment and manual measurements. Non-contact methods can have additional advantages since they are potentially scalable, portable, can be used for continuous measurements, and can be used on patients and workers with varying levels of dexterity and independence, from people with physical impairments, shop-floor workers to infants. In this work, we use face videos and the photoplethysmography (PPG) signals to extract relevant features and build a regression model that can predict pulse rate, and pulse rate variability, and a classification model that can predict force exertion levels of 0%, 50%, and 100% (representing rest, moderate effort, and high effort), thus providing a non-intrusive and scalable approach. Efficient feature extraction has resulted in high accuracy for both the models.