An attendance register to record students presence or participation in a lecture, tutorial or practical is usually done manually. This traditional attendance record system, where students have fill in their details on paper, is inefficient and students can cheat by getting their peers to sign on their behalf when they are absent for a lecture. This paper introduces a practical system for recording attendance automatically using facial recognition. The first part of the system is facial detection, which is achieved by using the Viola-Jones algorithm. The second part of the system, face recognition, is achieved through feature extraction and classification trained on the cropped Yale face database. A model is built by splitting 2452 samples from 42 people by a ratio of 3:1 for training and testing sets. Eigen faces are extracted using principal component analysis and are then fed into a support vector machine for training. The third part of the system uses the model to predict the student who enters the class and then records his/her attendance in a spreadsheet.


Download Final Thesis
Download Demo

Get In Touch.

Email Me:

Your message was sent, thank you!