The Application of Handwriting Recognition on Examination Cover Pages Using Artificial Neural Network

摘要

At present, the number of students is increasing while the number of instructors is insufficient in several colleges and universities. Some courses have the number of students participating more than 1,000 and the examination styles of these subjects are in form of short-answer or essay test. With these styles of tests, instructors have to give marks and collect scores for all questions and search names of students on a list for filling scores. These tasks are actually time-consuming for a number of students. To solve this problem, we invented an automated tool for detecting handwritten information on an examination cover page. Using the image processing and the artificial neural network techniques, all information on an image of the examination cover page was automatically extracted and then the summation score was calculated. High accuracies of 95.00% and 95.10% were obtained for letter and digit classifications, respectively. The processing time capacity of our tool and manual calculation were compared and the results showed that using the tool is faster with a factor of 2.41. This tool is useful and helps to reduce operating time for calculating the summation score, and exporting student information and the scores into a file.