Automatic system with artificial vision for grading multiple choice written tests for educational institutions ruled by the LOEI
Ricardo Jumbo-Herrera, Milton Labanda-Jaramillo · EasyChair preprint · 2018
This project develops a solution to the different problems that arise in the optical character recognition (OCR) in the field of education, specifically in the teaching area, focused on the recognition of characters in answer sheets, which are obtained from the application of written evaluations on a structured basis. To be implemented, Tesseract was used as the library for character recog- nition, for being robust, powerful and for the ability to be trained, also it uses leptony as an OCR algorithm; the main problems are shown when recognizing the answer sheets, as well as a series of methods for correcting characters and techniques that allowed obtaining a better result. Finally, results are shown among a sample of 100 students of an educational institution in which the effec- tiveness of the system was tested.