How and Why BriefCam Developed Proximity Identification
About Video Content Analytics
Countless businesses have video surveillance systems that record everyday interactions. In many cases, that infrastructure was installed primarily for security purposes, but by complementing existing investments with video analytics software, organizations are also leveraging their video content for business intelligence for a variety of use cases from traffic, demographics, navigational pathways, and much more. Powered by Artificial Intelligence and Deep Learning, BriefCam’s video analytics solution identifies, classifies, and indexes objects in video; the extracted and categorized data becomes searchable, actionable, and quantifiable and can be used for forensic investigation, alerting, and data visualization based on extensive filter combinations. By processing video in real-time and on-demand, users can benefit from increased situational awareness, as well as long-term insights into their environment
Applying Video Analytics to Overcome New Challenges
When news of the COVID-19 pandemic started to spread March 2020, we at BriefCam were presented with a new innovation challenge: we had to figure out how our video analytics technology could be suited — and further enhanced — to support efforts to curb the spread of COVID-19. Stakeholders from around the world, across our research and development and business strategy teams rapidly identified emerging and evolving market needs for the “new normal” of the pandemic world.
Our video analytics platform – with its extensive filters offering – was well positioned to meet those challenges. By combining existing analytic features and capabilities and developing new ones – including face mask detection and proximity identification – we were able to help support the ever emerging public health mandates. These new capabilities helped both end users and organizations encourage mask wearing and social distancing in real-time and assess overall and ongoing mandate compliance.
For real-time monitoring in a reopening office setting, BriefCam empowers employers to configure alerts for mask wearing and social distancing violations to be notified if staff or visitors are not practicing social distancing or wearing a face mask. Similarly, organizations can track compliance with mandates over time by leveraging BriefCam’s RESEARCH solution and aggregating data into various dashboards for monitoring and demonstrating compliance. This capability has empowered several hospitality venues, as well as corporate business entities, to safely reopen, give employees confidence to return, and maintain health and safety on an ongoing basis in their facilities or offices.
How does Proximity Identification Work?
To understand the geometrical layout of the scene in the camera view and map out the locations of the objects, our team refined BriefCam’s unique “scene geometry” learning approach which is how we effectively assign 3D distances to a 2D video scene. BriefCam’s software translates the camera view to a geometric X/Y plane so it can plot, detect, and extract the location of different objects. BriefCam does this by analyzing the scene’s makeup, including the “horizon”, and calculates the camera’s installation angle by analyzing the changes in the sizes of people as they move throughout the scene. The AI engine then transforms the scene into a top-down view, giving each person an X-Y coordinate which is then translated into real-world coordinates and distances (either in meters or feet, depending on the user’s settings). These coordinates and distances are measured between all of the objects in the scene once every .333 seconds for determining proximity violations. The proximity identification capability can thereby measure or calculate the proximity between the objects for a variety of key metrics, but the most common – right now at least – is the distance of six feet, or the social distancing standard for reducing COVID-19 disease transmission.
Proximity Identification for Contact Tracing
To enable contact tracing, proximity identification can be combined with other analytics, such as mask detection, appearance similarity, or face matching, so that operators can quickly search across multiple cameras and identify where and when a self-identified infected person interacted with others, and whether he or she was wearing a mask. Without compromising the anonymity of the infected individual, the operator can effectively determine which other people are at risk and recommend self-quarantine, while allowing unaffected individuals to continue coming to the office with confidence.
Preventing Crowding & Preserving Safety
While always a bit of a challenge to health and safety, crowding became especially problematic during the COVID pandemic, because it hindered social distancing efforts. BriefCam’s real-time alerting based on proximity identification and people counting enables security personnel to understand when and where crowds or queues are forming and proactively alleviate them through strategic decision making.
Ensuring that customers are not crowding in queues, aisles, terminals, or stations is also key to increasing customer satisfaction and positive customer/guest/visitor experiences. For example, in a setting such as a busy stadium, a shopping mall during the holiday season, or an airport during peak travel times, organizations can use proximity identification to assess developing crowds or bottleneck formations and determine whether a real-time response is needed.
Security teams can also use proximity identification to investigate claims of harassment or assault by checking the proximity of the accused with the accuser and validating whether any interactions took place. This can reduce litigation and expedite the process of validating a true claim or clearing a false one.
Monitoring Customer Engagement
In several environments, such as retail or banking, it is important to ensure that employees engage quickly and consistently with customers. Proximity identification can be used to evaluate whether sales or service managers are detected within a certain proximity of unrecognized visitors and to ensure that employees are proactively engaging with customers to maximize the customer experience.
BriefCam’s proximity analytics can also indicate whether customers or guests visit an information kiosk, advertisement, or product display, and how closely they interact with one. BriefCam dwell alerts can also measure how much time they spend in such a location. This is a key business intelligence data set that can guide advertisement, merchandizing, or layout decisions.
Realizing the Value of Video – During and After the Pandemic
BriefCam is very pleased that we were able to quickly adapt our technology to help companies and communities around the world overcome some of the challenges brought on by the COVID-19 pandemic. We are also proud that our comprehensive feature set continues to empower people, companies, and communities to improve workplace safety and optimize customer experiences and operational efficiencies, now and in the near post-COVID future.
Signup to receive a monthly blog digest.