Design and Development of Automated Student Coursework Mark Record System with OCR

 




 

Soh, Chin Huat (2016) Design and Development of Automated Student Coursework Mark Record System with OCR. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

[img] Text
Soh Chin Huat.pdf
Restricted to Registered users only

Download (1MB)

Abstract

The traditional method of recoding the coursework mark is by manually calculates the total coursework mark before it is being records to the system. This method is very time consuming because lecturers needed to mark over hundred or even over thousand students coursework in one semester. This project is to design and develop an automated student coursework mark record system with optical character recognition (OCR) to make the record process of coursework become fully automated. The system developed is able to generate a unique QR code which contains the students’ information such as student ID, student name as well as course title. The system generated QR code is printed together with the student information and also the table for the user to fill in the coursework mark by hand. Each section of mark allocated shall be input by the user before printing the coursework mark sheets. After the coursework mark is filled by the user, the coursework marks papers are input to an automated scanner, all the image is import into desktop or laptop of the user. After that, user able to extract the information a number of coursework mark sheet image in one time with the click of the recognition button. After the recognition, the student information and respective mark for each section of mark will be inserted automatically into database together with the calculated total marks. This project is designed by using the MATLAB R2015a software. The Microsoft Access 2010 is used as the system database to store the information of students together with the marks they get. The hardware used to implement the automated function of the coursework mark record system is Canon PIXMA E480 which build in with the automated document feeder (ADF) scanner. The main parts of this project are the automated record system development, encoding and decoding of QR code by import the java library into MATLAB, handwriting recognition by using OCR, and the connection of the database and the system. The automated system developed able to scan a number of papers from the scanner with output the image in many single files, hence the recognition system can process all the images which produced by automated scanner, and extract all the information in coursework mark sheet. Several experiments are carried out such as different accuracy of handwriting recognition for original and cropped image, the colour of paper used, and the colour of pen used. The result shows the cropped image has the high accuracy during handwriting recognition compare to the recognition of the whole paper. The overall accuracy of cropped image with blue colour ink and printed in white paper are 65.88%. Besides that, an experiment is carried out during decoding of QR Code shows that the decoding process does not depend to the paper colour. Hence, at the end of development, the project is able to record the coursework in just one click of button for decoding, handwriting recognition and database storing.

Item Type: Final Year Project
Subjects: Technology > Electrical engineering. Electronics engineering
Faculties: Faculty of Engineering and Technology > Bachelor of Engineering (Honours) Electrical and Electronics
Depositing User: Library Staff
Date Deposited: 22 Oct 2019 02:22
Last Modified: 22 Oct 2019 02:22
URI: https://eprints.tarc.edu.my/id/eprint/10872