OCR (Optical Character Recognition) is a line of research within image processing for which many techniques and methodologies have been developed. Set of pixels recognized based on the digitalized image and this study presents an iterative process that consists of five phases of the OCR. For this purpose, several image processing methods are applied, as well as two variable selection methods, and several supervised automated learning methods are explored. Among the classification models, those of deep learning stand out for their novelty and enormous potential.