HANDWRITTEN RECOGNITION ON ENVELOPES USING TESSERACT

Abstract

Optical Character Recognition (OCR) is the study of how to convert digital images captured or scanned from handwritten, typed, or printed documents into computer-understandable text. In the world, OCR technology has had profound impacts on many areas of production and life. The conversion of printed documents on paper into a compact and easy-to-search electronic form helps millions of pages of books and newspapers reach readers around the world. By combining with text-tospeech software, this document can be read aloud to people with visual impairments. Many post offices have applied an automatic mail sorting system based on envelope readers with OCR software installed. Banks read the contents of checks to combat money laundering, fraud, and even terrorism detection. OCR also enters everyday life through personal information devices (PDAs) that help users enter data by writing on the touch screen instead of carrying around a cumbersome keyboard. With the rapid development of technology, people want faster, convenient, and reliable tools that can meet their daily needs. This study is an endeavor to put forth a test process for handwriting recognition on envelopes using Tesseract Google's open source Optical Character Recognition (OCR) engine with Python. Moreover, the recognition problem is done with the support of OpenCV library. The handwritten on the envelopes will be captured by the camera. The process of building the identification system is done on the Windows operating system and the interface part of the system is done with the support of the Pyqt5 framework. The objective of this project is to extract the text from the scanned images, recognize the sender and receiver's addresses in it, display or store it in a database for further usage, and to enable a way in which processing of the documents will lead to eliminating the human touches and therefore dramatically reducing the processing time and the cost.