CNN based Digital alphanumeric archaeolinguistics apprehension for ancient script detection

Abstract

OCR is a most prominently used system in computer vision space. In the era of computer vision, the recognitiontechnologies are indeed evolved but there are still some difficulties for computers when reading handwritten text which can be resolved only after the introduction of machine learning. Recognition and verification of handwritten information is still a challenging problem in machine learning. Optical Character Recognition is a process of recognizing text or information present inside the images and converting it into a digital formatted text. Text recognition has immense applications in the academics, research, commercial and industrial fields. This paper is about an Optical Character Recognition for text recognition from the images which could be in any of the forms of handwritten text files as well as from the ancient manuscript (Language-English). This paper presents a novel machine learning approach to recognize the characters using CNN and the accuracy is found to be 73% approximately within a fraction of second. Later the recognized images are converted into the text file and then get translated into the preferable languages.