Video Recognition System Tracks Washing Patterns of Hospital Workers

A team from Stanford University built and tested a computer vision system that uses cameras to track clinicians that wash their hands, helps to identify offenders that don’t do it often enough, and hopefully lowers the rate of disease transmission through a hospital. If the approach proves itself, it may convince hospitals to introduce this technology on a wide scale as it may need only additional software to process the video that is already available through existing cameras.

The system is able to identify individuals and to recognize that they are probably washing their hands when right next to the washing stations. The software follows the clinicians as they move around the floor, tracking whether an infection has a chance to move from one place to another. Those individuals that pose the most danger of transmitting a bug around are identified and can be asked to improve their compliance.

A study of the system has been published to arXiv, the pre-peer review repository of new research, and it indicates that this approach may be one of the most effective and cheapest at managing hand washing compliance, not requiring special sensors or body-worn tags. Moreover, it can help to better understand the movement of people throughout a hospital ward and to improve the design and usage of clinical areas.