Online OCR (optical character recognition) using NVIDIA Jetson platform

• Main Author: Joshua Kurniawan
• Other Authors: Chau Lap Pui
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/74715
• Institution: Nanyang Technological University

This report documents the algorithm of Optical Character Recognition, to the purpose of detecting and recognizing the real-life text such as type word or scanned word. It details the process of the program to detect and recognize word from scene text images and explain the entire process of the program itself. Split into two major parts, the report explains within the Project set-up and on the latter part the Project process and development. The first part of the report deals with the set up to do the project, where multiple libraries is used and all of them is dependent to each other, thus comprehensive set up, and step by step guide is necessary to make library works properly. The second part is detailed explanation of the program processing the scene text images, so the program can detect and recognize the word within the images using OCRTesseract as the OCR engine, the report shows detailed explanation of the pre-processing needed to be able to minimize error of the program. Then it seeks to explain the process of feature extraction and text segmentation. Finally, the report explains the result of OCRTesseract when multiple variable in the images is change.