A machine vision system applied to the teaching of mathematics for blind or visually impaired children

This paper proposes the architecture and the procedure for the design and development of a new educational board game based on computer vision, which is an automated tutor oriented to special education institutions. For practical demonstration of the proposed system, we applied machine vision to educational board game focused on two-dimensional Braille math operations. Three tests were performed: 1) Test of performance of the machine vision based on cross-correlation algorithm, 2) Test of the algorithm that locates errors on the educational board game, and 3) Usability evaluation by visually impaired children and teachers at the Institute for Blind people Byron Eguiguren (Loja-Ecuador). As a result of 1), the average success rate is equal to 94.6%; as a result of 2), the algorithm locates 100% of possible errors; finally as a result of 3), aspects as interactivity, usability, likeability, and playability are met. These results indicate the high potential of the new application proposed of the machine vision systems; as a means for creating automated tutors applied to educational board games for visually impaired children, these teaching tools meet such characteristics as low-cost, fast and non-intrusive.