Using AI/Deep Learning to prevent retail inventory loss

Retail inventory loss (or “shrinkage”) is a serious problem, totaling about $100 billion annually—almost 1.8% of sales—worldwide. The issue is even more acute for those moving high-dollar goods, such as fashion and accessories. As traditional retailers grapple with ongoing market concentration, loss of market share to online sellers, and other pressures, there’s good news: Advanced analytics technology can help get a handle on shrinkage.

Of the four main causes of inventory shrinkage—shoplifting, return fraud, employee theft, and administrative error—theft by shoppers (or “external shrinkage”) accounts for the largest share (36.5% in the U.S. in 2017). This includes the growing occurrence of organized retail crime (ORC), in which groups of shoplifters operate strategically. An astonishing 94.6% of U.S. retailers surveyed in 2017 reported being victims of ORC in the prior year.

Shrinkage appears to be getting worse too: The number of retailers reporting more than 2% shrinkage annually grew from 17.1% in 2015 to 20% in 2018.

Not surprisingly, retailers are seeking solutions. While some are looking at a broad range of possibilities—facial recognition, motion tracking, reading emotions and gestures, and AR—most want to begin with solutions that are low cost and yield immediate returns.

Fortunately, there are effective technological solutions that satisfy these criteria. External shrinkage, for example, can be detected and prevented using inexpensive cameras and deep learning video analytics. At self-checkouts such a system can notify employees if an item passes a scanner without being scanned, stopping theft and deterring future attempts.

Based on its advanced real-time analytics platform, Megh Computing’s Video Analytics Solution addresses fraud prevention throughout the retail supply chain. It makes possible automated security measures to address all forms of inventory loss at self-checkout counters. The solution can be used to support inventory management systems to accurately identify items as they are checked in and out of unmanned distribution centers. Megh’s Video Analytics can also provide video surveillance to enhance physical security systems for access control for locations.

Megh Computing’s solution is based on a heterogeneous CPU+FPGA platform that can be deployed locally or in a data center. It maps the complete real-time analytics pipeline into the FPGA, including ingesting video streams, transforming video streams into resized image frames, and object detection and classification of image frames using deep learning systems. Megh Computing’s solution delivers increased throughput at low latencies with a >2x reduction in TCO.

Learn more at megh.com.

References

Loss Prevention: 4 Types of Retail Shrinkage and How to Prevent Them

Inventory Shrink Cost The US Retail Industry $46.8 Billion

Shrink cost retailers $100B last year

Organized Retail Crime: Television Drama or Retailer Reality?

National Retail Federation 2018 National Retail Security Survey

Machine Learning and the Store of the Future

Share this page
Share on facebook
Facebook
Share on google
Google+
Share on twitter
Twitter
Share on linkedin
LinkedIn

Megh Computing

Enabling the 3rd Wave of Computing in the Data Center