As losses from ORC and shoplifting continue to climb, large retail chains are on the hunt for new ways to secure their stores. One of the most popular new loss prevention methods that large retail chains are investing in is facial recognition, as evidenced by data showing that the compound annual growth rate (CAGR) for face recognition in the retail industry is a whopping 23.86%!
As large chains invest in facial recognition technology, it’s critical to keep in mind that not all face recognition systems are enterprise-friendly. Without a high degree of match accuracy and the right feature set, a facial recognition deployment likely won’t deliver the ROI that enterprises are looking for.
5 Enterprise Face Recognition Features
All facial recognition systems begin with assembling a database of persons of interest that customers’ faces can be matched against. A truly enterprise-ready facial recognition company can greatly expedite database creation by building a dynamic, curated watchlist of local, state and national persons of interest for customers.
An Airtight Matching Algorithm
A lot of facial matching algorithms just use tens or hundreds of feature detection points on a face in order to establish identity. But an enterprise face recognition solution should be using thousands of points on a face in order to establish identity. A better algorithm can result in a much higher degree of accuracy, which will undoubtedly result in a better ROI.
Enterprise facial recognition solutions have to be able to quickly scale flawlessly across hundreds (or even thousands) of locations. This means that you need to find a solution that is built to handle large deployments. The solution should also have a support team in place to handle installations, including optimizing cameras for lighting conditions and angle.
Built-in Privacy Protection
An enterprise face recognition company shouldn’t just think about its customers, but also its customers’ customers. That’s why it’s important to invest in a face detection system that is built to protect privacy. Some privacy-related features to look for include:
- Data encryption– Image data is encrypted both at rest and during transmission.
- Data breach precautions– Biometric templates stored within the facial recognition system should never be converted back into a face image in the case of a data breach.
- Data purging– Surveillance data should be automatically purged at regular intervals in accordance with industry best practices
- Anti-profiling– Facial recognition systems should be built to prevent profiling by race, age, gender or national origin.
The right set of analytics can help large enterprises prevent future crimes. First, it’s crucial for enterprises to be able to view aggregate data across multiple store locations. Face recognition analytics should help companies easily identify which times of day and which locations are experiencing the highest number of matches, so that each location can be accurately staffed up. Enterprise face detection systems should also offer historical data about when criminals or banned customers have visited locations in the past. Finally, the system should offer the ability to segment face recognition data into groups, zones, locations and other tags in order to better identify retail crime patterns.
Looking for an enterprise face recognition solution? Contact our sales team today to set up your free demo!