Optical character recognition (OCR) on images using template-matching and image correlation

Zakari, Nur Hakimah (2020) Optical character recognition (OCR) on images using template-matching and image correlation. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Optical character recognition (OCR) on images using template-matching and image correlation.pdf - Submitted Version Download (427kB)
	Text (Full Text) Optical character recognition (OCR) on images using template-matching and image correlation.pdf - Submitted Version Restricted to Repository staff only Download (2MB)

Abstract

Throughout the year, Optical Character Recognition (OCR) has become a demand as the high usage in images that converted images into editable machine-coded text within multimedia or digital field. Optical Character Recognition (OCR) techniques is a process to extract the text by improving the quality of the digital image. OCR system includes Pre-processing, Segmentation, Feature Extraction and Classification. For classification, template matching and correlation used to identify the font and extract the text. For this template, 6 fonts have been chosen which are Cambria, Lucida Sans, Dubai Medium, Microsoft YaHei, Malgun Gothic and Lato through 4 various sample to match the template and analyses the accuracy based on the correct recognition (%). This OCR system will be implied in MATLAB software with Image Processing Toolbox. The purpose of this project is to obtain the text within the images that contain crucial important information where the text will be output through notepad

Actions (login required)

View Item