The School of Computer Science is pleased to present…

Webcam-based Optical Mark Reader for Grading Assessments 

MSc Thesis Defense by: Muhammad Fahad Zafar 

Date: Wednesday, January 18, 2023

Time:  02:30 PM to 03:30 PM

Location: Essex Hall, Room 120 

Reminder: Completion of QR-link and sign-in sheet attendance is mandatory.

Abstract:  

Optical Mark Recognition (OMR) is a process of retrieving information from the forms where users put marks in particular locations. OMR technique has been improved significantly since its inception, moving from specialized hardware-based solutions to more adaptable and sophisticated software-based solutions using a scanner. Recently, webcam-based software solutions have emerged because of their affordability and ease of use. In this thesis, we propose an improved webcam-based OMR solution for grading the student tests given on a specifically formatted answer sheet. We apply various image processing techniques to achieve effective grading, resolving technical problems stemming from multiple factors like image orientation, skewness, and uneven illumination. In particular, we propose a new variant of the Otsu binarization technique for image segmentation to effectively separate the background and foreground. The proposed grading solution aims at maximizing the criteria for acceptable operations so that 100% accurate grading among those accepted ones can be achieved. The evaluation of the proposed method is two-folds: (i) the proposed variant of Otsu will be compared with other binarization algorithms for its effectiveness in background separation; (ii) the grading solution will be compared with the state-of-the-art webcam-based solutions where the individual contribution of each applied image processing technique to the grading enhancement will be identified.

Keywords: Optical Mark Recognition, Image Processing, OpenCV, Otsu Binarization

MSc Thesis Committee:  

Internal Reader: Dr. Imran Ahmad

External Reader: Dr. Fazle Baki

Advisor: Dr. Jessica Chen

Chair: Dr. Dan Wu