In the morning you walk into a local convenience store and immediately hear your name broadcast over the public announcement system welcoming you; you grab the drink you want and just go without checking since the AI system has debited your purchase already; and you never have to worry about waiting in lines when you go shopping. No lines, no checkouts, no registers-- this is not the dream fantasy about what shopping would look like in the future, but a real ‘’intelligent supermarket’’ experimental platform created by the Tianjin University Machine Learning and Data Mining team and the TJU Zhongke Vision Extension Artificial Intelligence joint lab. The application of artificial intelligence technologies is turning smart retail into a reality.
Target Detection leads to immediate payment.
Target detection is an artificial intelligence technology that can automatically extract and identify multiple targets in real-time processing in complex scenes. Tianjin University research group successfully achieved smart retail settlement through target detection technology.
In the TJU ‘’smart supermarket’’ test platform, a single product can be settled immediately once it had been taken off the counter for consumers to
pay through their smart phone using target detection technology.
The team also put forward a new design of ‘’vision+QR” self-settlement code for those who buy a large amount of groceries. People only have to put the items on the scanner and target detection and recognition technology can quickly do the payment. The entire experience is quick, efficient and painless.
Facial recognition makes intelligent supermarkets consumer-friendly.
Facial recognition is a biometric technology which works on the basis of the feature information of human faces. Using this technology, the team from Tianjin University has made a breakthrough that they could detect and recognize the faces of customers in “intelligent supermarkets” by means of scanning faces, inputting information, filing and constructing the corresponding “memories”. These memories are fused with the intelligent supermarket system serving as the supermarket’s “brain”. Owing to this technology, intelligent supermarkets can analyze whether customers are regular customers, VIPs or return customers, and even surprisingly recommend goods based on their buying preferences and habits. This makes the supermarkets more consumer-friendly.
Pedestrian Detection enables supermarkets to capture what customers need.
Pedestrian detection is an intelligent technique to determine whether there is a pedestrian and achieve a precise mapping in an image or video sequence by computer vision. Assisted with the techniques such as pedestrian tracking and recognition, it can be applied in artificial intelligence systems, vehicle assisted driving systems, intelligent robots, intelligent video surveillance, human behavior analysis and intelligent transportation, etc.
Having installed an intelligent monitoring system in the intelligent supermarket, the research team from Tianjin University can better grasp customers’ shopping preferences and shopping patterns in accordance with the statistics and analysis of passenger flow, thus providing effective data for shelf display. This facilitates the work of making the optimal layout scheme in different periods, in a bid to greatly enhance shelf space utility rate and user comfort. In the team’s scheme, the supermarket has been endowed with the ability to “think” and turned into a smart supermarket that could capture what customers need during intelligent retailing. Zhu Pengfei, a young teacher in charge of the experimental platform, said that there are some technical bottlenecks in computer vision and machine learning used in intelligent supermarkets at present. The establishment of the experimental platform serves as a perfect training ground for basic research and students' innovation and entrepreneurship, thereby actively promoting the grounding and popularization of artificial intelligence.
Currently, the campus intelligent convenience store, provided by the machine learning and data mining team of Tianjin University and the Joint Laboratory of Artificial Intelligence of Tianjin University-SeetaTech, is in operation on the Peiyangyuan Campus, in which students can get an easy access to the intelligent service and buy goods conveniently without going off campus.
By Gao Han & Gu Junru
Editors：Eva Yin & Doris Harrington
