A system for automated attendance through face verification, as well as finding interest levels of a student in e-learning
Principal Investigator: Dr. Aurobinda Routray
Co-researchers: Shubhobrata Bhattacharya
Objectives
Algorithm to classify the mark the attendance of a student through face verification
Estimation of attention index through face image sequences using deep learning
Estimation of expression index through face image sequences using deep learning
Estimation of head pose index through face image sequences using deep learning
Motivation
Attendance in most large classrooms is conducted manually, resulting in considerate amount of wastage of time
Chance of impersonation is more in cases of manual attendance
There are no standard systems to provide the interest level of the student in a course
Algorithms
The student needs to register for the course with Name, Frontal face images, and other details
Upon login-in, the student needs to verify for his face against the credentials to mark attendance as well as verify impersonation
There is provision to play a video lecture, either online or offline recorded
With the start of a lecture, the recording of facial images of the student begins
The recording of facial images is analysed framewise
Each facial frame is check with a classifier concerning expression, eye state and head pose
Finally the indices are computed