image

Computer Vision for Retail

What do face-detection cameras and self driving cars have in common? They all apply some form of Computer Vision (CV), a scientific discipline that enables machines to make useful decisions about real physical objects and scenes based on images. How Computer Vision is transforming retail?

Early research on Computer Vision started over 50 years ago, and its application across industries have grown along with our understanding of the discipline. Most digital cameras today recognize faces in a picture, OCR software in scanners converts scanned documents to text, and vision-based biometrics also famously helped identify an Afghan girl by her iris patterns.

Some early applications of Computer Vision in retail come from e-commerce, but increasingly, it is being used in physical retail stores to perfect shelf merchandising, enhance operational efficiencies and create a frictionless experience for shoppers.

Here are few inventive ways brands and retailers are using computer vision to bring innovation.

Blurring the line between online and offline

Sometimes you see something you want to buy, but have no information about it. In that case, a tool called Lens can help. It’s been launched by the photo sharing website Pinterest as a beta product, and it could aid the in-store experience. Most of the large Ecommerce websites like amazon use the same technology to help users find a product using photos.

Facial and behaviour recognition

Gourmet candy retailer Lolli & Pops uses facial recognition to identify loyalty members as they walk into the store. Computer Vision then enables a personalized shopping experience: by scouring their purchasing history and preferences, the system can make personalized product recommendations specific to each shopper.

By treating them as individuals – and more importantly, as VIPs – the system instills brand loyalty, and converts occasional shoppers into regular customers. Both of which are good for business.

Digitize Shelfs

The beauty and simplicity of Computer Vision is its ability to turn actual images into actionable insights in order to help brands and retailers focus on fundamentals in the store. By “digitizing the shelf”, companies now get real-time situational awareness about what’s happening on the shelf. The directives range from the obvious (such as: “go to the back room and get a box of product to fill an empty space”) to the more sublime, such as instructions to reduce the number of products of the same type that are sitting side by side of a competitor and increase your own products by that same amount.

Non-mobile users get role-based insights on a huge array of retail metrics that tell them exactly what’s happening on the shelf and what to do to ensure the best shopping experience and drive better sales.

Seamless Checkouts

Computer Vision can also help when it comes to one of the worst parts of the shopping experience: queuing for the checkout.

The Amazon Go concept store in Seattle tracks shoppers using CV, with sensors on the shelves detecting when they pick up an item. It then registers all the items in the shopper’s shopping basket with the Go mobile app, and does away with the checkout process altogether – the shopper simply leaves the shop, with the Go app taking the money automatically from the shopper’s nominated credit card. The receipt is sent straight to the app.

The ever-connected shopper experiencing frictionless retail is truly where we’re headed, made possible by a combination of Computer Vision and deep learning.

Wondering how real shelf images are turned into actionable analytics? Check out how we can help the retail industry.