Abstract

The cornerstone of a quality education system is Outcome-Based Education (OBE), which revolves around predefined outcomes or goals. Based on the best practices of NEP 2020, all the education institutions and accreditation authorities rely on Outcome-based teaching, learning, and evaluation. To measure the defined outcome of each course, it is essential to enter the question-wise marks in the student exam portal for CO attainment calculation. Traditional methods adopt the manual entry of marks. This is time-consuming and error-prone, leading to additional work for each examiner after evaluation. This project uses computer vision and AI to automate the identification and calculation of marks from the corrected handwritten exam answer sheets. It extracts information like roll numbers, question numbers, and marks, calculates total marks, and exports data to Excel sheets, hence reduces manual effort and improves efficiency. Deep neural network-based pre-trained models are used for the effective identification of handwritten marks. We have trained and tested the proposed system through continuous assessment and internal examination answer sheets available in our institution.

Authors

J. Annrose, A. Shibin Sam, S. Samnath, Jennings Rousseau Raj, I.R. Shanjeya Bhaarath
St. Xavier’s Catholic College of Engineering, India

Keywords

Handwritten Digit Recognition, Outcome Based Education (OBE), Measuring Outcomes, Attainment Calculation, Computer Vision, Deep Learning