Systems Outage Information: Click here to visit

MSc Thesis Defense Announcement of Rahul Bharadwaj Raveendran:"Effective Auto-Grading with Webcam"

Tuesday, May 24, 2022 - 13:00 to 14:30


The School of Computer Science is pleased to present… 

MSc Thesis Defense by: Rahul Bharadwaj Raveendran 

Date: Tuesday May 24, 2022 
Time:  1:00 pm – 2:30 pm 
Passcode: If interested in attending this event, contact the Graduate Secretary at with sufficient notice before the event to obtain the passcode


Optical Mark Recognition (OMR) refers to the technique of electronically extracting data from the marked fields on the printed forms. Stemming from a hardware solution with dedicated equipment, it has gradually been replaced with emerging software technology and portable scanners. In particular, the webcam-based software solutions to OMR turn out to be more attractive due to their high affordability and flexibility. In this thesis, we propose an efficient and effective webcam-based OMR solution for auto-grading student work. The efficiency is achieved by maximally dropping the requirements on the operations of the end-users. The effectiveness is understood together with the handling of the technical difficulties originated from the factors like the distance between the grading sheet and the webcam, the rotation and orientation of the grading sheet, the brightness of the environment, etc. Various present OpenCV image processing techniques are combined to accurately identify the region of interest and the reference points on the grading sheet. This is followed by a specific algorithm to detect the orientation of the grading sheet. The proposed solution tries to maximize the acceptable operations and aims at providing 100% accurate grading among those accepted ones. This is followed by a specific algorithm to detect the orientation of the grading sheet. The ultimate goal of this thesis is to maximize the acceptable user's operations and to provide 100% accurate grading among those accepted ones. We have successfully extended the perspective tilt of the orientation of the OMR sheet to the range of [50, 90] degrees. Within this range, our experiment shows that the proposed algorithm can reach 100% accuracy in retrieving the region of interest and in grading the tests. Compared to the existing methods in the literature, ours provides a 7% performance improvement in terms of detecting the region of interest in the OMR sheets. 
Keywords: Image processing, Optical Mark Recognition, OpenCV 

MSc Thesis Committee:  

Internal Reader:   Dr. Ikjot Saini  
External Reader:  Dr. Balakumar Balasingam        
Advisor:                Dr. Jessica Chen 
Chair:                    Dr. Xiaobu Yuan

MSc Thesis Defense Announcement

5113 Lambton Tower 401 Sunset Ave. Windsor ON, N9B 3P4 (519) 253-3000 Ext. 3716