The core technology behind TurboX featured at the exhibition is the object recognition technology from Thundersoft. A technology-driven company, Thundersoft has invested heavily in OS and technological development since its inception. Imaging algorithm represents one of the core technological strengths of the company. Employing its proprietary Camera processing and optimizing technologies, Thundersoft object recognition offers an attractive solution for pre-processing.
Object recognition is computer vision technology that identifies the wanted object in an image or an array of graphics. The object must be identified despite many different forms and sizes, or even in spinning, moving conditions or partially hidden.
The aim of Thundersoft-S.POINT cooperation is to implement object recognition in self-help payment. Building on its success of over 1,000 innovative implementations, S.POINT opts for the combination of PC, USB transferred-data, and installed devices to implement the system.
In the TurboX solution, Thundersoft provides smart algorithm based on Qualcomm chips and custom-made OS tailored for retail scenarios. The strong expertise of Thundersoft addresses the challenge whereby the embedded algorithm cannot work with cloud-based data.
The recognition device built on image recognition module and sensing area is able to implement precision data collection. The successful recognition of the bread is done on three dimensions:
• Object features (bread with sesame, cheese, and sugar power, etc)
• Form factor
In addition, S.POINT found in tests that lighting would affect recognition accuracy. Therefore, the sensing area lighting is complemented with fill light, thus compensating for the lighting conditions in the shop.
When it comes to the physical device, the recognition device presents a display. The consumer sees the scan result on the display when attempting for self-help payment. The scan window can scan QR codes attached to goods such as juice, milk, and sandwiches. The consumer selects the payment method and makes the payment for the bread as well as other goods.
S.POINT offers a complete workflow based on the specifications of object recognition and self-help payment system:
1. Put the object in the sensing area
2. The system identifies the object by using the camera above
3. The identified object is shown on the display
4. Confirming the scan result, the consumer may make the payment online
5. Slip printed after the payment
The above technology can also be used for inventory. A complete database complementing central kitchen inventory helps reduce costs and prevent waste. For the consumer, the self-help payment process helps save queuing time in case of heavy in-store traffic. The store benefit from reduced crowdedness. The easy interface by S.POINT ensures a smooth workflow.
The object recognition-enabled TurboX, jointly developed by S.POINT and Thundersoft (Northern Light VC tech partner), was presented at CES Asia 2018 in June, 2018.