Smart attendance system via facial recognition using Tensorflow Facenet model / Ahmad Muizzuddin Shahrel

• Main Author: Shahrel, Ahmad Muizzuddin
• Format: Thesis
• Language: English
• Published: 2020
• Subjects: Programming languages (Electronic computers); C (Computer program language). C++; Detectors. Sensors. Sensor networks; Pattern recognition systems
• Online Access: http://ir.uitm.edu.my/id/eprint/35667/1/35667.pdf; http://ir.uitm.edu.my/id/eprint/35667/

Student absenteeism problems are often a major issue for every school and are often associated with low student scoring status. This can be attributed to the weakness of the current method of taking attendance which used a sheet of paper. The system is developed to improve the current attendance system where several problems appear during the school attendance taken. For example, the student deceived their attendance, time consuming for waiting to sign the attendance and the lack of ability to track the absence of many students. This system can give better support and help the teacher’s work on the attendance. Popularity of biometric recognition where human unique physical body as a measurement access control for identification and authorization has led to the development of Smart Attendance System using facial recognition. An identification system is used to digitally verify a person by comparing each known face and information corresponding to the database. The Tensorflow Facenet model is the technique used for the recognition where it's called as one shot model where directly learning and mapping the face images into Euclidean space where distances are used to calculate the similarity of faces. With the facial recognition feature, this system can easily manage and automate taking attendance and safely record it in the database.