The namesake of our project as a whole, Facial Image Recognition Neural Network, is our current focus. In it's current iteration FIRNN is based around improving registration and attendance systems for universities. Paper or ID card systems are the norm in the majority of universities currently.
Paper systems are not only a waste of resources but also rather unreliable. There are multiple points of failure during the process of collecting attendance for lectures; the lecturer could forget to bring the register to the lecture, a student may forget to write their name down (or may have messy handwriting), the lecturer may then forget to collect the register at the end of the lecture, and then mistakes may be made when entering the register into the database on the computer.
-
FIRNN aims to replace current registration and attendance systems in as many universities in the UK as possible by 2030.
-
The network is based on the ResFace architecture and as such has been trained on a vast set of data. The model has over 200 network layers, with many varying stages of feature extraction.