Asia Express - Data Communications
Taiwan Unveils New AI-Enabled Visual Checkout System to Help Reduce Queue Time By 50%
July 30, 2020

The Industrial Development Bureau (IDB) of Taiwan’s Ministry of Economic Affairs (MoEA) has assisted Viscovery, a Computer Vision AI startup, in developing the AI-enabled visual checkout system, called “Vision AI Checkout”, to solve the problems in retail like reducing queuing time and training costs, according to MIC (Market Intelligence & Consulting Institute), a government think tank and research institute in Taiwan.
Viscovery has cooperated with Taiwanese POS terminal manufacturers such as Posiflex and Firich Enterprises and industrial PC (IPC) manufacturers such as Advantech and Axiomtek to launch a total solution. The Vision AI Checkout can be easily integrated into any point of point-of-sale (POS) or self-check system with an application program interface (API). Instead of hiring a cashier, the AI-enabled visual checkout system is able to detect and recognize baked goods such as bread, cakes, and pastries with the accuracy rate as high as 98%. Combined with edge computing technology, the system can facilitate self-checkouts and help cashiers to reduce checkout time by more than 50% (from 60 secs to 1 sec), and ultimately enhance operational efficiency and in-store customer experience. For instance, Taiwan’s bakery IJySheng and several bakery businesses have adopted Viscovery’s "Vision AI Checkout". 85°C Bakery Cafe has also adopted such system in its chain stores in the United States.
According to the MIC, the adoption of AI-enabled visual checkout systems using computer vision and machine learning technologies for innovative applications in retail is anticipated to grow faster than those using traditional barcode and RFID scanning technologies. Thus far, Amazon, Standard Cognition, Trigo Vision, AiFi, Sensei, Zippin, and Grabango have unveiled their own AI checkout solutions that come with AI lenses to be installed on the ceiling and/or the shelf of stores.
Taking the variations in store designs and the overall implementation cost into account, "AI Point-of-Sales (POS) terminal" and "AI shopping cart" are two key areas of focus for retailers, according to MIC. They not only can help reduce the complexity of recognition process but also greatly narrow down the scope of monitoring from the entire store to small blocks to reduce deployment costs. Diebold Nixdorf, Mashgin, and Tiliter are major suppliers of AI-enabled POS terminals while AI-powered shopping carts are mainly provided by Amazon, Caper, and Imagr.