AnyVision Visitor Analytics
Facial recognition technology can be used to visually track visitors in a public space (e.g., a retail store or mall) and build a detailed and integrated view of what they did, that can then be mined, for example, for retail optimization (store layout, use of digital signage, etc).
AnyVision is a well-funded, young, Israeli company that has leveraged image processing and machine learning technologies to build a sophisticated facial recognition and person tracking system (BetterTomorrow) that has found good traction in security applications (like casino visitor identification and tracking). AnyVision Insights is an adapted version of the security platform that can be used for other visitor analytics purposes.
Business Need: Bring to a “bricks-and-mortar” store as much of the visitor analytics and understanding as possible that has been developed in e-Commerce (e.g., via click tracking analytics) and used for innovative on-line retail optimization.
Traditional retailers are in an existential struggle with their e-Commerce competitors (e.g., Amazon) and need to do everything possible to maximize the value of an in-store experience. Knowing what visitors do in-store, and optimizing based on that data is obviously foundational.
Edge Need: These applications are computationally intensive (leverage GPU accelerators) and bandwidth and latency sensitive. Edge resources are an attractive alternative to requiring and having to operate in-store servers.
Ease of Incorporation: Data-intensive measurement and analytics is native to e-Commerce, but disruptive to most traditional retailing (although clearly needed). Given the common use of security cameras in retail, the physical infrastructure is largely in place. The biggest impediment is the public fear of facial recognition (Insight doesn’t require knowing who the person is, but it will be tempting to track other visits and integrate with other multi-channel marketing information such as actual purchases and/or on-line activity).