Master Thesis: Edge Computing For Biometric Access Control Based On 3D Antispoofing And DL Recognition

This is the official work thesis I published for my Master Degree in Computer and Automation Engineering. The project prototype has been developed autonomously and tested over real life examples.


The main idea behind this work thesis is to design and build a versatile access control system which not only provides control access mechanisms through RGB cameras, but also implements a 3D verification process to distinguish between real individuals and fake 2D images/videos. To accomplish this, the system employs advanced Intel© RealSenseTM Cameras which provide access to 3D data through Active Stereo/Laser Scanning technology. In addition, the project also aims at implementing a deeper level of granularity to check if a single individual (after being recognized) is allowed to access an area or to unlock a device at a given time, on a given day. This ”second level” control reinforces the overall system, resulting in stricter and more reliable access security measures.

Description