Retrofittable Apartment Access Device
Leveraging Facial Recognition

By Andrew Palmer

adc

Abstract:
The aim of this design project was to explore the facial recognition capabilities of embedded
and compute resource limited devices, such as the Raspberry Pi. Through this exploration, an application
was developed using the facial recognition features of the OpenCV library as its pivotal subsystem. This
design project’s application was a retrofittable facial recognition enabled apartment access device. By
analyzing a user’s face, the device allows or prohibits entry into one’s home or apartment. Additionally,
the device also hosted a web server that allowed a user to grant access without needing facial recognition.
When access was granted from either the facial recognition application or the web server, the device
transmitted a radio frequency message to a secondary device attached to the user’s buzzer. The buzzer
was then activated, which opened the door. Using this application, a user could register common visitors
and delivery services, which could alleviate the frustration of missed deliveries from not being at the
buzzer. Additionally, the facial recognition application was implemented in Python and C++, which
offered comparison between the two implementations on various metrics.

Report (pdf)

Code (zip)