Handwritten Character Recognition for Automated Evaluation of Malaysian Primary Schoolers� Handwriting Readability

Abstract

Examinations are no longer obligatory for Lower Primary School children in Malaysia as they now would be assessed with PBD, a more objective classroom-based assessment. While it may reduce issues such as peer pressure and depression among children, however, PBD would also impose high workload to the teachers, which in itself could be another reason for less objective and inconsistent evaluation due to stress. Therefore, this study aims to explore the in-depth aspect of handwritten character recognition technique that will be able to evaluate the readability of students handwriting based on the criteria established for a writing assessment. Results obtained from preliminary data analysis show that contour based feature, statistical features, structural features, and global transformation and moments features sets that are the techniques with the ability to extract all five (5) features of a handwritten word which match the criteria that measures handwriting readability that are size, space, slant, line quality, and formations.