TY  - BOOK
AU  - Milad,Ali Ali Alhadar
AU  - Yurtkan,Kamil
TI  - AUTOMATIC CLASS ATTENDANCE MANAGEMENT SYSTEM USING FACIAL SYNTHESIS AND RECOGNITION / 
PY  - 2022///
KW  - Human face recognition (Computer science)
KW  - Dissertations, Academic
KW  - Support vector machines
N1  - Thesis (PhD) - Cyprus International University. Institute of Graduate Studies and Research Management Information Systems Department; Includes bibliography (sheets 124-135)
N2  - ABSTRACT
Classroom attendance systems have been a principal part of the learning process of
conventional classroom institutions. Many processes have been devised over the
years for the adequate recording of student attendance in classrooms to ensure
students attend their classes which is synonymous to good academic performance in
students. The rise of the digital age in the 21st century has enabled the introduction
of several electronic technological platforms for classroom attendance purposes, one
of such technologies is the use of human face recognition for classroom attendance.
Unfortunately classroom attendance has been faced by several challenges which
many studies have tried to mitigate and proffer solutions to. This study proposed a
methodology to mitigate the challenges of varying face poses and facial expressions
which both face recognition-based classroom attendance systems have been reported
to face over the years. This research leverages the use of 3D analysis of facial
expressions to synthesize face image poses and facial expressions from a single input
image of a subject. The study uses six basic facial expressions which it implements
using a 3D model of set of points triangle mesh points connected in 3D space of 20
geometric feature points on human faces. The synthesized facial images are applied
Local Binary Pattern Histogram (LBPH) texture operator on, to enable the extraction
of histograms which are used to train a Support Vector Machine (SVM) classifier.
The study uses two different human face databases; FEI and BU-3DFE, the two
databases are used to evaluate the proposed methodology separately, with the FEI
database images yielding an average of 71.78% performance accuracy and the BU-
3DFE database yielding an average of 78.57% performance accuracy. This designed
methodology is further developed and applied as a recognition system for classroom
attendance scenarios which led to the achievement of a satisfactory performance
accuracy for individual subjects tested on it.
Keywords: Classroom Attendance System, 3D Face Analysis, Face Recognition,
Support Vector Machine, Local Binary Pattern Histogram
ER  -