Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
final year thesis final year thesis
Typology: Thesis
1 / 62
This page cannot be seen from the preview
Don't miss anything!
Automatic Exam Corrector
Date Description Author Comments
The following Software Requirements Specification has been accepted and approved by the following:
Signature Printed Name Title Date
Automatic Exam Corrector
Automatic Exam Corrector
Automatic Exam Corrector
We declare that this project is our original work and has not been presented for a degree in any other university.
Name Signature Date
This project documentation has been submitted for examination with my approval as university advisor:
Project Advisor : Mr. Wondimagegn Desta _________________________
Automatic Exam Corrector
Accuracy is most important thing for this project. The project has to be able to identify what the user has written on the paper in order to give mark based on the answer inserted by the teachers or the one who prepared the questions. Since we are going to apply machine learning technique for text recognition of the student’s handwriting, we need to gather large data that can help us prepare a good model that can precisely recognize the students’ Amharic and English handwriting. Therefore, we need to gather a satisfying dataset and this data gathering will be the most important part of our project phase. In addition to that, for the android mobile application, users need to have a mobile device with good camera quality so that they can take a vivid photoshoot of the student’s answer sheet for the application to recognize the handwriting.
2.1 Product Perspective Currently MoE is using the Optical Mark Reader (OMR) for marking national examination which can be as extremely large as hundred thousands of student examination answer sheets and the answer sheets have their own format which the machine would not mark correctly without that proper paper format. The main and obvious limitation of the machine is that it won’t read handwriting text which makes it difficult to automate exams given in other formats and these other formats are common in every school and university. Often times, these format restrictions makes it difficult to automate exam correction if not in national levels with a huge budget correcting machines. But our proposed Automated Exam Corrector will resolve all these problems which makes it a convenient way of handling schoolwide as well as nationwide examinations. Moreover since it's going to use printers and Jet scanners with conventional desktop computers, which most of the time already exist in the school compounds and teaching institutes, it makes it cost wise, affordable and consequently pragmatic solution.
This system will provide a primary input method for any Academic institutes from high school to universities and National & Regional Examination center. Moreover, it will reduce the human power needed for this activities. Specifically, the system will have the following functionalities. The system should recognize the students answer including the student basic information. The system should allow the examiner to export the result in to different format (csv, web portal...) and gives statically analysis. The system should support both Amharic and English answer sheets. The system should be able to print student results on the answer sheets.
Automatic Exam Corrector
There are four types of user that interact with the system.
The most crucial constraint of this project is related to the accuracy of the system and its output. The software we design is related to exams, which needs accurate correction process and if not maintained can lead to product unreliability, and even more to project failure. The system is expected to have at least the same average accuracy of a human corrector in order to be usable.
There are three assumption that we think for the proper function of the system
The first assumption is that the system will generate standard answer sheet. So anyone who wants to use this system have to generate the answer sheet template first.
The second assumption is that the desktop system would have access to a scanner with windows operating system installed on the working computer. This would not apply for the mobile application system.
The third assumption is that the mobile application system would have access to a good camera with Android operating system installed on the working mobile. This would not apply for the desktop system.
Automatic Exam Corrector
3.1.3 Software Interfaces
The Desktop application with Scanner and Printer integration requires a minimum of Windows 10 Operating system since the framework used to develop the application is the UWP (Universal Windows Platform). For the device integrations of Printer and Scanner, respective device drivers are required to be acquired from the vendor and be pre-installed on the system.
3.1.4 Communications Interfaces
The basic communications required by the system are communication between the application with the printer and scanner. Both communications happen through the device driver interfaces so there is no additional communication interface required.
Table 1: Functional Requirement-
Name FR-01: Browse Answer Sheet Images
Description This functionality enables the user to browse already taken/scanned images of answer sheets from a file explorer.
Inputs Inputs to this functionality is a set of already scanned image files which can be browsed and viewed.
Source Filesystem
Outputs The output of this feature is the file location (URI) of the images to be processed.
Destination Answer Corrector component
References See UC-01 for detailed steps
Table 2: Functional Requirement-
Name FR-02: Scan Answer Sheet Images
Description This feature enables the user to integrate to a scanner and start scanning task to image answer sheets fed to the scanner.
Inputs The scanning feature requires an integrated scanner device which is fed with answer sheets to be analyzed.
Automatic Exam Corrector
Source Scanner driver
Outputs This feature generates digital images of the scanned sheets to be processed by proceeding features.
Destination Filesystem and Answer Corrector component
References See UC-02 for detailed steps
Table 3: Functional Requirement-
Name FR-03: Correct Answer Sheet
Description This feature is the most significant feature of the system which powered by Artificial Intelligence to analyze answer sheet regions, recognize characters and do comparisons to determine whether answers are right or wrong.
Inputs Inputs to this functionality is a set of scanned images, answer types on the sheet and preset answers to the questions.
Source FR-01 or FR-02 outputs and user
Outputs The output is a set of results specifying whether an answer is correct or not.
Destination Persistent storage(local or remote database)
References See UC-03 for detailed steps
Table 4: Functional Requirement-
Name FR-04: Generate Check Mark Grid
Description This feature generates check mark(R X) grids and exam format to be printed.
Inputs Input to these feature is a set of answers and their size and format.
Source User
Outputs An output of this system is a transparent image grid that resembles the
Automatic Exam Corrector
Table 7: Functional Requirement-
Name FR-07: Export Exam Report Grid
Description This feature exports every examinee results to spreadsheets, PDF and the optionally to Databases for mass student access through Web portal.
Inputs Input to the system is a formatted report representing the exam schedule.
Source Persistent storage(local or remote database)
Outputs Output of this feature is a document of different file types which can be viewed and printed.
Destination Filesystem
References See UC-07 for detailed steps
Table 8: Functional Requirement-
Name FR-08: Generate Web Portal Template
Description This feature will generate a bootstrap Web portal template for students to view their exam scores. This template can be customized and maintained by the enterprise as per their requirement.
Inputs Input to this feature is exam results and scores of all examinee.
Source Persistent storage(local or remote database)
Outputs An output of this feature is a basic Web portal bound to a database which can be accessed by Web protocol.
Destination Web server
References See UC-08 for detailed steps
Automatic Exam Corrector
Actors The following are actors
**1. Scanner
Table 9: UC-01: Browse Answer Sheet Images
Name UC-01: Browse Answer Sheet Images
Goal Browse and specify already found images from directory
Actor User
Precondition There is already captured answer sheet
Success Post condition Specified images by the user will be selected for further processing
Failure Post condition Show failure message with its related cause
Trigger Browse command UI
Table 10: UC-01/Browse Answer Sheet Images
Main Success Scenario: UC-01/Browse Answer Sheet Images
Extension Scenario: UC-01/Browse Answer Sheet Images
Automatic Exam Corrector
Alternative Scenario: UC-01/Browse Answer Sheet Images
None
Table 13: UC-03: Correct Answer Sheet
Name UC-03: Correct Answer Sheet
Goal Take an answer sheet image and correct based on preset answers
Actor System
Precondition Compatible answers are set properly for corresponding answer sheet.
Success Post condition Answer sheet correction results will be stored in the system including all related problems arising from the process.
Failure Post condition Show failure message with its related cause
Trigger Picture captured or Correct command UI
Table 14: UC-03/Correct Answer Sheet Images
Main Success Scenario: UC-03/Correct Answer Sheet Images
Automatic Exam Corrector
Extension Scenario: UC-03/Correct Answer Sheet Images
Alternative Scenario: UC-03/Correct Answer Sheet Images
Table 15: UC-04: Generate Check Mark Grid
Name UC-04: Generate Check Mark Grid
Goal Generate printable exam answer sheet
Actor User
Precondition N/A
