Supporting automated grading of formative multiple choice exams by introducing student identifier matrices

We propose a solution which enables automated attribution of formative exams to students. We define the student identifier matrix, which is a graphical representation of a student ID number. The matrix is annotated by the student during the exam. We introduce a robust image processing procedure which enables automated recovery of the student ID from the matrix. We show that the proposed procedure achieves a recognition rate of 100% on 296 samples in presence of skewing, rotation and offsets introduced by the printing and the scanning processes.