System with Optical Mark Recognition Based on Artificial Vision for the Processing of Multiple Selection Tests in School Competitions

Buleje C.Y.C.; Atencio Y.T.P.; Tinoco E.E.C. · Proceedings - 2020 46th Latin American Computing Conference, CLEI 2020 (2020) · DOI: 10.1109/clei52000.2020.00027

Resumen

The present study originated with the need to automate the process of qualification of multiple selection tests in school competitions at a low cost, this was done using the technology of optical mark recognition (OMR) with artificial vision algorithms for answer cards fill circles type and with conventional scanning devices, it was possible to determine the level of precision based on four important issues: economic cards, detection and correction of cards with amendments, student coding and time reading, the results show an accuracy 97.21% for the primary level and 99.24% for the secondary level that allowed to create images with the qualified tests taking into account the correct, incorrect and blank answers and printing the respective score, also the detection and storage of errors in reading was made due to printing and scanning errors, this solution allows time saving considerable since it was achieved a qualification time of 0.83 seconds for the primary level and 0.65 seconds for the secondary level, the present solution will allow the educational institutions that organize school competitions to be able to access this technology together with an integral registration system.

Instituciones: Universidad Nacional José María Arguedas