FaceTag (2019)
GitHub
Facial Recognition Attendance System
The project is based on FaceNet (2015) which claims 99.63% accuracy on the Labelled Faces in The Wild Dataset. The face detection and model training algorithm was developed by Wut Hmone Hnin Hliang which is my supervisor at the time.
Designed and developed the user interface complete with database for an Ubuntu application. Data collection, data cleanup and an ideal training set are built for 35 TM R&D staff. Test reports are produced for further improvements
The user interface is developed using Glade and GladeGTK with Python. The database is developed with mysqldb library for python.
Conclusion :
This system is only to be implemented in low security settings as it only extracts the facial features in a 2D plane which can be easily manipulated using pictures.
The accuracy is astounding as improvements and research have been conducted for deep learning and facial recognition














