Devanagari manuscript labeling using feature extraction: A survey and future directions

Translation of multiple images or documents into editable text, that can be decomposed is termed as OCR. Different automated machine learning tools have been used in the literature that segregate the data into characters. This automatic segregation helps to convert the copies into electronic format. This paper deals with multiple feature extraction techniques and also the data labeling tools that have been used extensively for character recognition. This paper presents the present state of art techniques to provide further direction of research for the researchers.

I. Introduction

Optical character recognition (OCR) - a fully grown, advancing, energizing, and helpful field that serves as an establishment stone for the some advanced applications like speech recognition, medical field, bio-metric recognition, and many more. Optical Character Recognition (OCR) coverts the paper based documents into editable formats. Instead of scrolling the huge files for searching a particular piece of text there arose a need of conversion. The technique used to convert the raw formats into computer understandable language is known as OCR.