July 22, 2019
Facial recognition technology has come on leap and bounds in the last five to ten years, primarily due to the plummeting costs of high-resolution imagery. However, the rise of machine learning has helped create accurate, repeatable and scalable solutions that can perform superhuman tasks.
Rather than staff staring at multiple CCTV feeds, or putting bouncers on doors, real time alerts whenever unwanted visitors enter a site can provide significant reductions in staffing costs and improves security of both patrons and staff. The combination of machine learning and AI facial recognition algorithms helps to protect the respectability of the business and ultimately reduces loss of revenue by preventing cheats and hooligans from entering, safeguarding the experience for the business’s desired customers.
Amazon have recently taken this technology even further with their new Amazon GO stores in major US cities, that are completely cashless and don’t use manned or self-service tills. Once you’ve downloaded the app and entered a store, Amazon’s extensive camera system tracks customers and logs what they put in their baskets. After customers have finished shopping, they simply have to walk out the store and Amazon will charge the items to their account. This completely revolutionises the traditional ‘Retail Customer Journey’ by making the experience far more convenient for customers, whilst reducing staff requirements.
In addition, the metadata gathered around popular products is informing where Amazon place high-margin items and could provide more actionable data than traditional market survey methods. This can improve revenue and return custom, as Amazon finds their feet in brick-and-mortar retail.
By examining length of strides, wobble, speed and various other factors, Amazon have found that gait tracking is a more accurate way to follow their patrons.
Gait tracking is improving safety in city centres too, where behaviour can be inferred by how someone stands or walks. Imagine a high-value store that has cameras outside that can detect when someone is casing the joint and acting ‘shady’. The cameras can identify the threat, report it to the police, and thereby become a proactive public safety tool.
Given that the principle of Minority report was that three clairvoyants could see into the future and identify crimes before they happened, the use of AI-powered tracking could start turning that fiction into a reality.