The Effectiveness of Image Preprocessing on Digital Handwritten Scripts Recognition with The Implementation of OCR Tesseract

摘要

Optical Character Recognition (OCR) has been widely discussed in various topics in the rise of robotics, artificial intelligence and computer vision. OCR has become a solution in extracting characters from the image into machine-encoded text. This research aims to discuss character recognition from digital handwritten image. However, characters recognition problems using OCR has been more or less solved. OCR mainly implemented in reading characters from scanned of printed documents. In this research, image preprocessing including convert to grayscale, morphological operations and noise removal has been successfully boost the accuracy score of OCR performance. The average success outcome resulted to 79.26% in reading characters from the image.