Employee Attendance on Raspberry Pi using Facial Recognition

This project is done during SDAIA Academy Bootcamp as a part of METIS Kaplan course requirements in collaboration with Abdulrahman Alrubaiya & Mohammed Alghamedi

Abstract

With all the rapid changes in modern world, facial recognition been an integral part in almost all domains. From safety and security to industrialization, that concept grew largerly; in addition, utilizing that to identify objects and people in a quickly manner is a must in a fast growing economy. Many applications can take advantage of object detection. In our project, we are trying to record employee attendence using facial recognition on Raspberry Pi microcontroller.

Problem

The current pandemic gave us new ways to leverage technology. Existing biometerics the likes of fingerprint requires contact with the device. This would raise a challenge in preventing the spread of the disease since surfaces may be contaminated with the virus. In our project, a contactless approach is what we look for; therefore, an object detection/facial recognition is what we aim to achieve.

Data and models used

Our data is going to be combined with transfer learning and pre-trained models. GoogleNet and other several models would be incorporated to determine if a specific employee is counted present in the attendance system or an unknown alien is inside the organization premise.

Tools and hardware

• Raspberry Pi 4
• Logitech USB webcam
• Numpy
• OpenCV
• PyTorch (or Keras)
• scikit-learn (Keras wrapper if decided on Keras)
• Matplotlib
• Seaborn
• Twilio - Communication API (SMS)