By Stephanie Weagle
Mitigating the Spread of COVID-19
Across many industries— from retail to manufacturing and transportation —the COVID-19 pandemic has created a new set of workplace health and safety practices that aim to prevent the spread of the contagion among employees and customers. Hospitals and nursing homes, among others, must embrace these new practices, because many of them come into contact with COVID patients and employ employees who may carry the virus and are at high risk of contracting the illness.
As we can see looking back at media coverage from April 2020 – such as Time Magazine and NPR – from the onset of the pandemic, many public health experts have agreed that contact tracing, improved hygiene and sanitation, mask wearing, and physical distancing are the pillars of the fight against COVID-19. Contact tracing enables others – healthcare workers among them – to learn if they have had contact with an infected person and to proactively isolate themselves to prevent further risk to their family, friends, and co-workers, and to take action quickly if they have any symptoms. Of all the public health and safety recommendations, contact tracing is especially difficult to implement, because it requires the cooperation of so many people. From contacting a diagnosed individual to confirm he/she has contracted COVID-19 to working with the person to review their movements and interactions, healthcare workers and contact tracers work tirelessly to map out those at risk as comprehensively as possible and must ultimately reach out to exposed individuals to advise them to monitor themselves for symptoms and self-quarantine for 14 days.
Facilities that treat COVID patients have an even greater need to conduct effective contact tracing among their employees and model best contact tracing practices for other organizations.
In some ways, hospitals are uniquely qualified and positioned to be the backbone of contact tracing efforts, because some people who test positive for the virus end up receiving care in a hospital setting, where they can be easily identified. But to break the chain of transmission on a healthcare campus and in the surrounding communities, those healthcare settings will need a combination of human effort and technology.
Intelligent Video Surveillance as a Natural Contact Tracing Driver
While it is vital to quarantine those who have tested positive for Coronavirus, it is often challenging to retrace their interactions during the contagious period. Depending on one’s daily routine, a person may interact with dozens of patients, colleagues, or visitors and not be able to fully recall every point of contact. Asking a staff person (especially one who is seriously ill) to speak with a human contact tracer to describe all of his or her personal interactions will likely prove difficult, incomplete, and expensive.
Cell phone apps are just one of various technology solutions that are enabling contact tracing; thus far, however, such apps have not proven to be very reliable, especially because there is no mass adoption to support their effectivity.
Most hospital campuses already have a network of video cameras across their premises, in parking lots, gift shops, restrooms, cafeterias, storage rooms, and, of course, diagnostic and treatment facilities. By pairing existing video surveillance infrastructure with video content analytics technology, hospitals can conduct contact tracing much more effectively. Powered by artificial intelligence (AI) and deep learning, video analytics software transforms live or recorded video from video monitoring systems into structured metadata that can be searched, analyzed and acted upon. The technology can detect all the objects in a scene, then extract, identify, classify, and index them to enable a variety of analytic applications. Contact tracing with intelligent video surveillance is enabled through face recognition, appearance similarity, and proximity identification.
Facial Recognition & Appearance Similarity
Face recognition extracts unique identity features from a digital face image and compares them against a database or watchlist identified persons of interest. Significant deep learning research and development have enabled facial recognition algorithms to effectively identify possible match faces in scenarios where humans could not, so that human operators can assess and validate potential matches.
If an employee self-identifies as testing positive for COVID-19, with permission, the employer can upload his/her photo to the video analytics system to enable a surveillance operator to quickly search and filter video footage for appearances of that individual across all surveillance cameras on a hospital campus over the course of the contagious period. In this way, the facility can identify other employees and hospital wards at risk of contracting the virus. Operators can also use appearance similarity filters – broader person attributes or characteristics, such as the person’s gender or color of his/her clothing – to focus the video search for relevant results.
Proximity Identification & Face Mask Detection
While it’s critical to identify the people and places with whom the individual interacted, another equally helpful metric to assess risk is by determining whether the now diagnosed person was observing physical distancing and face mask wearing recommendations when the contact occurred. Intelligent video surveillance can be used to determine the proximity between the infected person and others across days and cameras and whether either person was wearing a protective mask. The operator can pre-set safe distances and time durations that, when violated, indicate higher risk of virus transmission. By filtering video based on these parameters, healthcare organizations can further narrow their search for potentially affected people. At risk employees, patients or visitors can be notified that they have had interaction with an anonymous person confirmed to have COVID-19, and the hospital can recommend next steps for mitigating further virus spread and ensuring the wellbeing of the at risk individual.
Preventing COVID Transmission with Video Analytics
Beyond contact tracing, video analytics also offers solutions for proactive prevention. Hospital campuses can apply the technology to monitor or enforce other COVID-19 mandates, including mask wearing and physical distancing. For example, a video content analytics system can be configured to alert on face mask violations in real-time, so that staff can proactively ask individual visitors to comply with on-site regulations. Rule-based alerts can also be used to detect and monitor crowding or physical distancing violations – when individuals are detected standing in close proximity for long periods of time, pre-set by the user. Of course, as with all alerting solutions, a critical role in the detection and identification process is the assessment and decision making of the human operator, who can effectively use the information made available by the technology to determine the potential risks and appropriate responses.
Using people counting analytics, intelligent video surveillance operators can monitor occupancy in indoor and outdoor spaces to prevent crowding and to optimize maintenance operations. By detecting the number of people that enter particular spaces, such as restrooms, operations managers can ensure that the campus is cleaned based on quantifiable traffic data instead of arbitrary scheduling. The system can generate a maintenance alert after a pre-configured number of people have walked across a restroom or food court entrance and, over time, the facility can evaluate long term traffic data to plan cleaning schedules based on usage and traffic in these areas.
Of course, intelligent video surveillance has been a worthwhile investment and driver of critical healthcare facility applications long before the Pandemic started – and will continue to maximize investments in video surveillance long after concerns over COVID-19 recede. Various hospital business departments can use video content analytics for improving situational awareness and accelerating post-incident investigations by on-site police and security teams; assessing aggregated demographic data about hospital visitors for operations teams to improve on-site services and retail offerings; and evaluating pedestrian and vehicle traffic patterns to drive strategic planning around infrastructural renovations or construction.
Stephanie Weagle is the Chief Marketing Officer at BriefCam, the industry’s leading provider of Video Synopsis® and Deep Learning solutions for rapid video review and search, face recognition, real-time alerting and quantitative video insights. Stephanie leads the company’s global marketing initiatives, accelerating market adoption of BriefCam’s comprehensive video analytics solutions. Before joining BriefCam, Stephanie was Vice President of Marketing for Corero Network Security, where she led global marketing for the company’s cyber-threat mitigation product portfolio. Previously, Stephanie held senior marketing roles at Lionbridge Technologies and Novell, Inc.