A Comparative study of GOCR, Tesseract and Improved Tesseract for Character Recognition

Abstract

Optical Character Recognition is a technique which is used to change printed document or handwritten character to readable text format. If the text is in printed form against a complex background, it is usually very difficult to perform the detection method. OCR localization approach is used to detect the horizontally aligned text automatically. Tesseract and GOCR are most widely used tools in optical character recognition to recognize the characters from the given input. Tesseract provides better accuracy with low error rate however, in some cases it is unable to recognize text written in image, so there is a need for improvisation of Tesseract tool. This paper introduces a new approach to improvise Tesseract tool by using Artificial Neural Network