In this project, we are going to emulate an automated passenger onboarding kiosk used in airports. The process will be conducted by passengers that look into a camera for the 30s to collect their facial details, and use a "scanner" to scan their Passport and their boarding pass. From this data, the following information will be extracted.
- Emotion
- Identity
- Name and Date of birth
- Personal info and flight specific info.
To tackle this project Azure services will be used. Different azure services will be used to extract data from the different sources. There will be 3 different sources of data. Video, Id, and Boarding pass. Azure computer vision, Face services, and video analyzer will be used to extract and match a passenger's face, furthermore extracting passengers' emotion. While Azure form Recognizer will be used to extract information from identification cards and boarding passes. There will be also a task to build an Azure cognitive service-based computer vision for identifying lighter on the carry-on luggage of passengers. The process of validating the data collected is as described below:
- Boarding Pass and ID: The text data collected will be cross-checked to check its validity. The flight details are presented to the user on the screen
- ID and Video: The id and video faces will be cross-matched to validate the passenger's identity
- Video: Collect the emotion of the passengers
- Other: Identifying a lighter in passengers luggage
The End Result and summary of the resources to be used is shown below. The passanger will be prompt a message with the process being succesfuel see picture below. The threshold for confidence level will differ, for face validation 80%, lighter detection 50%, text information 80%. As lighter is a critical thing flagging detection is important as passing the detection can be dangerous. Face and text detection is set high to avoid false detection leading to false verification.
