5 most inventive uses of Computer Vision in retail
What do smile-detection cameras, barcode scanners and autonomous vehicles 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 is this technology being put to use in the world of retail?
Early research on Computer Vision started over 50 years ago, and its application across industries have grown in sophistication 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 five inventive ways brands and retailers are using vision-powered innovation.
- Context is king: blurring the line between in-store and online
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.
It works by recognizing the object and providing contextual information about it. Perhaps it’ll tell you who designed the piece of furniture in what year, or it’ll suggest other clothing that’ll go with a certain pair of shoes. Just take a photo of the item, and the app does the rest.
- Facial recognition: identifying regulars, rewarding loyalty
Gourmet candy retailer Lolli & Pops uses facial recognition to identify loyalty members as they walk into the store. CV 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.
- Transform the health of every shelf, in every store
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 facings (how many products of the same type that are sitting side by side) of a competitor and increase your own facings 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 shelf and what to do to ensure the best shopping experience and drive better sales.
- Shopper measurement: Analyzing footfall, pass-by traffic, interactions and more
Aurora by RetailNext is the first sensor designed specifically to meet retail’s complex needs. It counts in-store footfall traffic like so many other sensors of its ilk, but also adds texture to the data – it includes the capture rate of pass-by traffic, and breaks down shoppers’ paths around the store. That way, you can see which promotions capture engagement, and which turn customers off.
But it doesn’t just monitor the shoppers. It also adds customer and associate interaction, providing real-time visibility into in-store service engagement. Plus it can be used to drive personalized marketing and messaging campaigns.
- The frictionless store experience: an end to checkouts?
CV 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 this ‘under the hood’ video to see the Trax Computer Vision platform in action.
