Exploiting CCTV Cameras for Hand Hygiene Recognition in ICU

The monitoring of hand hygiene activities can effectively reduce infection and contamination in the Intensive Care Unit (ICU). In this paper, we created a clinical dataset using CCTV cameras installed in ICU to explore the feasibility of recognizing the hand-washing steps of clinicians. A video processing architecture including hand landmark detection and classification is presented. Such a system can potentially be used to alarm clinicians to follow the guidelines of hand hygiene. The experimental results show that the average accuracy of our methodology can achieve 95% under the personalized model and 56% under the generalized model. The preliminary results suggest that hand hygiene is subject-dependent, which is related to individuals’ palm size and washing habits. The cross-subject modeling or subject-adaptive learning can be applied to further improve the accuracy of recognition towards a more generalized solution. The insights of the study are helpful for designing a hand hygiene scoring and alarming system as a part of hospital IoT. The hospital data and code are available at https://github.com/SunnySideUp11/Hand-Hygiene-ICU.