A lightweight, real-time application to invigilate an exam automatically without requiring any manual invigilator, expensive hardware or data-set training. Suited for users with lesser bandwidth.
Click on the picture below to see embedded video :-
If the above link is somehow not working - Click here
During the run, following suspicious images were found that were saved for the future reference
1. Python 2.7.12+
2. OpenCV 2.4.*
3. numpy 1.11.0+
-
For online exams, the existing applications are heavy on user's bandwidth since they require continous video streaming.
-
Since suspicious images would be compressed & saved into some directory, they may be used for future reference by instructor too.
-
Usual Eye-gaze tracker softwares require user to NOT to be in bright enough lighting. Using Contrast optimisations, this application is somewhat receptive to that.
A major defect in a lot of existing proctoring softwares is :-
1. Majority of them requires good enough webcam for proper functioning.
2. Some require a dataset for training the model at first - Training the dataset in real-time actually slows down the application.
3. Most of them require a physical proctor at some remote location keeping an eye on students constantly. This application automates it with good enough accuracy.