AWS for Industries

Seeing dollar signs: Ways to leverage computer vision in retail stores

Artificial intelligence (AI) has been around since the 1970s. However, only in the last 10 years has AI-powered computer vision become a mainstream reality, thanks to cloud computing. Here are several scenarios in which computer vision can enhance the retail shopping experience and improve store operations.

Cashierless checkout

The most obvious use of computer vision in retail is a checkout-free or cashierless checkout experience, like Amazon’s Just Walk Out solution. It’s the technology that powers Amazon Go stores allowing customers to select the items they want and… just walk out of the store. The cameras mounted throughout the store can discern when a consumer takes an item from the shelf or puts the item back. The AI-based solution charges the items to the customer’s credit card via a smartphone app. This is quite literally frictionless retailing, and it sets the bar for consumer expectations—no checkout required.

Supporting new safety protocols

In our pandemic world, we’ve learned that wearing masks and social distancing can help prevent the spread of infectious diseases. AWS Partners  ViTech Lab and SoftServe are using computer vision technologies to make indoor shopping safer with mask detection and shopper counting solutions that alert retailers when rules are broken. Amazon is even providing real-time social distancing feedback in warehouses. This technology helps consumers and employees feel safe as they resume normal activities.

Visual search

Have you ever had difficulty describing a particular product to a store clerk? Often, it’s easier to show a picture of the item to the employee. Product discovery technologies are a rapidly evolving use case for computer vision thanks to companies like ViSenze, which provides a solution to many other retailers, including Uniqlo, Rakuten, and Zalora. Today, ViSenze can allow retailers to access visual data from the store to forecast sales performance of in-store display items. AI and machine learning are the foundation of image recognition and product recommendation technology solutions.

Preventing accidents

Injuries in stores, either to shoppers or employees, can be a challenge for retailers. Cameras are a good way to prevent many potentially dangerous situations, such as spills or smoking at gas pumps. Badger Technologies put robots in Giant and Stop & Shop grocery stores to patrol aisles “looking” for hazards. When an issue is detected, the robot lights up to alert store personnel of the situation.

Stock validation

Auchan, a supermarket chain in Portugal, used AWS Partner Trax’s “Retail Watch” store monitoring solution to detect empty shelves. The solution helped the company reduce replenishment times to a single day. The same solution also verifies shelf pricing, which is normally a repetitive, manual process. This solution helped Auchan increase on-shelf availability by 3%, reduce price anomalies by 75%, and save 250 labor hours.

The computer vision field is ablaze with so many new solutions that allow store associates to focus on customers instead of repetitive, manual tasks, which lowers costs, speeds reactions, and generally improves and elevates the customer experience. Fortunately, today the tools exist to develop solutions that incorporate cloud-native AI capabilities into vision systems in distributed store environments. AWS built services such as Amazon Rekognition, AWS DeepLens, and Amazon Kinesis Video Streams to help deliver these types of solutions cost-effectively. Retailers should consider an infrastructure that supports these and future scenarios in their stores. We’re just scratching the surface today, and many more new ideas are emerging.

David Dorf

David Dorf

David Dorf leads Worldwide Retail Solutions at AWS, where he develops retail-specific solutions and assists retailers with innovation. Before joining AWS, David developed retail technology solutions at Infor Retail, Oracle Retail, 360Commerce, Circuit City, AMF Bowling, and Schlumberger’s retail and banking division. David spent several years working with NRF-ARTS on technology standards, is on the advisory board for the MACH Alliance, and supports the Retail Orphan Initiative charity. He holds degrees from Virginia Tech and Penn State.