Skip to content

Home

Jump to bottom
jackinsjh edited this page Jun 29, 2020 · 22 revisions

Presentation Youtube Link

https://youtu.be/Y7ib7HPLcys

Developers

Motivation

An automatic scoring project has been launched to help professors, teachers, part-timers and assistant teachers.

Brief Description of Project: Overall flow

Details of Overall Flow

First of all, there is the same test paper that many students took. And then there's the answer sheet, and there's the metadata that's processed the answer sheet. If you combine this with the algorithm, all the students' scores will come out.

Brief Description of Project: Technologies

Brief Description of Project: How to process picture

  1. Enter the number of questions
  2. Input unmarked test paper
  3. Designate 4 edges of the paper by clicking it
  4. Paper normalized
  5. Input each problem’s metadata
  6. Designate each problem choice’s area by dragging,After that, press corresponding key to determine whether it must be checked or unchecked.
  7. Calculate the difference between normalized marked/unmarked paper and grade!
  8. (option) Show the picture of marking so that the grader can grade descriptive problems
  9. Figure out the result

Used Techniques: OpenCV

  1. Resize
  2. Get clicked / Dragged coordinates
  3. Grayscale
  4. GaussianBlur
  5. Compare marked/unmarked papaers (SSIM index comparison)
  6. Normalize difference image by threshold

Used Techniques: QtDesigner

  1. Generate the '.ui' extension file
  2. Convert to '.py' extension file
  3. Merge the '.py' extension file (front-end) with back-end file.

Used Techniques: pytesseract

  1. Resize
  2. Grayscale
  3. GaussianBlur (Smoothing)
  4. Result of text file

Program Architecture

Documentation

Go to see the details... https://github.com/jackinsjh/Automatic_Scoring_Program/wiki/Documentation

Clone this wiki locally