|Customers: S7 Airlines Siberia Airline
Contractors: VisionLabsProduct: VisionLabs Luna
Project date: 2017/03 - 2019/11
On July 3, 2019 the S7 Airlines announced the beginning of system testing of face recognition at Domodedovo Airport. The technology is started the business hall in a zone of a departure of internal runs where passengers can get without production of documents. In the future "passengers will be able to enter from an input the airport building before boarding, having never got the boarding pass", the press service of carrier reports.
The software of Luna from VisionLabs company which is based on the trained neural networks allowing to identify the passenger and to personalize services was unrolled. It is noted that formation of the database of photos happens only from the personal consent of each passenger.
Anyone having the ticket of a business class can take part in face recognition system testing. For this purpose the passenger needs to approach on front desks of S7 Airlines 104, 105 or 106. There the employee of airline will accept consent to processing of biometric data and will take the picture of the client of airline.
After that the passenger will be able to pass to the airport business hall, without showing the boarding pass on an input. His face will be recognized by cameras, and administrators will receive the notification with a name and a surname of the passenger, the flight number and also information on access to the business hall. Cameras are also integrated into the screens installed in the business hall on which the board of runs is displayed.
According to S7 Airlines, the sensing technology of persons also allows employees of the business hall to see which of travelers left in time the business hall and who needs to remind of the coming landing. After completion of testing other business halls will be connected to a system, and the list of the personalized services constantly will be replenished, speak in the company.
At the beginning of 2018 it became known that S7 Airlines, the participant of global aviation alliance oneworld, is going to integrate a sensing technology of persons (face recognition) into work of business halls at the Moscow Domodedovo Airport. The software on the basis of the trained neural networks will allow to identify passengers of S7 Airlines and to personalize services for visitors of business halls.
Is planned that the video cameras installed on an input to business halls will "recognize" passengers of S7 Airlines and also to trace which of travelers in time left the business hall and who needs to remind of the coming landing to run. The system based on technology of "computer vision" integrated into a TV-panel with the schedule of departures will show to the passenger information on that run on which it is registered.
"We actively use the most modern technologies for ensuring comfort of passengers at each stage of a travel. In particular, the progressive sensing technology of persons will allow us to offer passengers the personalized services in business halls in waiting time of a departure. For example, it will be possible to obtain up-to-date information about the status of run and any changes, having just taken from a table the tablet with "computer vision", – Svetlana Kulyukina, the director of the department of passenger experience of S7 Group notes.
The sensing technology of persons is based on a possibility of the trained neural networks to independently develop and apply an algorithm of comparison of objects. A system transforms the seen person to a set of digits then compares it to own database. At the same time, thanks to the high level of data protection, inverse transformation from a set of digits in the photo is impossible.
In December, 2017 together with VisionLabs company testing of a sensing technology of persons at office S7 Airlines was held. In entrance group of the building video cameras which among all passing through tourniquets "recognized" those employees whose photos were loaded into a system were installed. For testing difficult conditions were intentionally selected: weak illumination of the premises and a noncooperative operation mode of a system when for recognition the person does not need to look specially in the camera, and is enough to pass by it. As a result at a test stage a system made about 50,000 photos, the accuracy of recognition was about 90%.
Formation of the database of photos will happen only from the personal consent of each passenger of S7 Airlines. The airline considers different options of data collection, including proprietary mobile application of S7 and photographing when scanning the boarding pass on an input to business halls. Photos of passengers, as well as all personal information which is processed by S7 Airlines it will be protected according to the legislation.