Skip to main content

Self-Checkout Kiosks Streamline In-Store Shopping

Computer vision, ai, artificial intelligence, digital display

Brick-and-mortar stores are still a thriving market segment. But for most retailers, competition is on the rise and profit margins are being squeezed. They need solutions that enhance customer experiences while at the same time lowering operational costs.

Innovative IoT technologies are helping in-store shopping experiences catch up with those online. And smart, self-service options are providing innovative ways to create better customer experiences while managing store operations. For instance, retail stores such as Amazon Go are showing the way with self-serve and cashless payments.

With that in mind, NexCOBOT, a subsidiary of the global IoT solutions company NEXCOM, sought to address these challenges and take advantage of new opportunities.

“Retailers have pain points such as high labor expenses. And a lot of the time, they can't even find employees to staff the store,” says Nelson Chang, Manager of the NexCOBOT Automated Retail Department. “The purpose of our solution is to help them address these challenges by creating a unique shopping experience.”

Watch this short video to see self-checkout in action (Video 1).

Video 1. NexCOBOT uses AI systems and learning algorithms to create a unique in-store shopping experience. (Source: NEXCOM)

AI and Deep Learning Make for Smarter Service Kiosks

NexCOBOT’s Smart Self-Checkout Solution uses facial recognition, object recognition, AI, and machine learning to streamline purchases from member login to cashless checkout. There’s no need to scan items one by one or go through the hassle of looking up products when a barcode is not available. As a result, customers have a better shopping experience while store owners reduce employee overhead costs.

The solution is powered by an Intel® Core processor, using object detection and facial recognition to streamline customer checkout. The AIR AI PC is the engine that drives the solution and can render multiple high-resolution display outputs simultaneously.

The Kiosk itself is equipped with cameras, an interactive video touch screen, RFID reader, bar code scanner, and receipt printer (Figure 1).

The Smart Self-Checkout Kiosk solution performs tasks like recognizing specific products.
Figure 1. The Smart Self-Checkout Kiosk solution performs tasks like recognizing specific products. (Source: NEXCOM)

NexCOBOT uses AI and deep learning to pre-train the system to recognize specific customer products. For example, a bakery may sell hundreds of different types of baked goods. The system will use photos from every angle of each product for object training.

And every item, from cookies to sliced bread, each can require more than 300 images to be correctly recognized. Behind the scenes, NexCOBOT uses a deep learning model to automatically generate 500 images from just 50 photos.

“We use the Intel® OpenVINO Toolkit as the inference engine,” says Chang. “It helps us to accurately recognize more objects in less time.”

The solution also provides store owners with data analytics via its Smart Retail Dashboard. A range of information can be viewed in real time on a digital display in the back office. Data such as transaction volume, popular items, hourly revenue, inventory status, and more enables store owners to make better business decisions.

RFP Ready

The Smart Self-Checkout Kiosk is available as an approved Intel® RFP Ready Kit (RRK) designed to help integrators achieve faster time to market.

The RRK makes it easier to deploy a proof-of-concept (PoC) that allows prospective customers to experience the solution in action. The kit includes all the hardware, software, tools, and support necessary to implement an edge-to-cloud solution right out of the box.

Faster Time to Market Through Partnerships

Developing and deploying smart retail systems like the NexCOBOT solution can be a significant challenge for solution integrators, ISVs, and OEMs. Self-checkout systems must be customized for different types of stores and products. But retailers remain cautious about investing in new technologies.

“Even in the initial staging, a customer may take three months to make the decision to invest in a PoC,” says Chang. But by leveraging the NexCOBOT RRK solution, retail SIs don’t have to build AI models from scratch or design and manufacture hardware kiosks.

For integrators working in the retail segment, self-service checkout offers new opportunities. And NexCOBOT offers a way to pursue these opportunities, with an RRK that can be scaled to support a customer’s entire retail environment and products.

But still, store owners may hesitate to pay for even a proof-of-concept deployment. This is another area where NexCOBOT can help. “System integrators trying to win new retail business are challenged by their customers’ budget constraints—especially for new technologies,” says Chang. “The value of our partnership with Intel® goes beyond technology. As a member of the Intel® IoT Solutions Alliance, we sometimes get help in funding a Self-Checkout RRK PoC, which makes the whole process move much faster.”

Ultimately, retailers want to increase revenue and drive down costs. To do so, they need to become more relevant to consumers. This means both removing obstacles for customers and improving business operations.

They’re looking for ways to provide a differentiated value proposition to consumers, and self-checkout solutions is certainly one way to do so. And the NexCOBOT RRK gives integrators a leg up to benefit from new opportunities in the retail segment.

 

This article was edited by Christina Cardoza, Associate Editorial Director for insight.tech.

About the Author

Asli Bilgin is a senior executive with over 20 years of experience in strategic business development, sales, consulting, software development, and management. She runs her own management consultancy business and serves as an author for Pearson Education. Asli is a leader in STEM community affairs and outreach to empower girls and women in technology, collaborating with MIT and Tufts University.

Profile Photo of Asli Bilgin