jin-zhe / OMR_Tools

Some useful tools for Optical Mark Recognition (OMR)

OMR_Tools

Some useful tools for Optical Mark Recognition (OMR). Note that only Python 3 is supported. There is no backward compatibility for Python 2 so you if you'd like to use Python 2, you'll have to make your own modifications.

Installation

First, get this repo and its submodules on your machine:

git clone --recursive [e-mail]:jin-zhe/OMR_Tools.git

I recommend installing the necessary python packages via Conda which has become the de facto virtual environment manager for AI/ML projects (which this might potentially turn into). Conda also makes the installation process really simple because it is also a package-management system. If you rather not use Conda, you may refer to OMRChecker's original installation guide using pip. Should you do that, just make sure you have the important packages listed under Packages.

Installing with Conda

1. Follow this quickstart guide to install and understand how to use Conda
2. Create a new Conda environment with necessary packages:

• conda create -n omr_tools python=3.5 opencv matplotlib numpy pandas tqdm poppler -y

3. Enter newly created environment

• conda activate omr_tools

4. pip install pdf2image imutils (these are the only two you have to install via pip)

Packages

This repo has been tested to work with the following python package versions:

Package	Version
opencv	v3.4.2
opencv-contrib-python	4.2.0.32
matplotlib	3.3.4
numpy	1.19.2
pandas	1.2.3
imutils	0.5.4
tqdm	4.56.0
pdf2image	1.14.0
poppler	0.81.0
PyPDF2	1.26.0

opencv-contrib-python

Submodules

As submodules are included in this repository, make sure you get the right one. If you didn't clone this repo with the --recursive flag, you'll need to run:

git submodule update --init --recursive

OMRChecker

• Branch: extension-framework
• Commit: 'bc9310135602bfedf47c0db99ab60c5aae9935c4'

---

Scripts

stuID_checker.py

Conducts OMR on student exam script scans to determine their student ID.

Sample usage

python stuID_checker.py -i $script_dir -o output.csv -r

Arguments

Longform	Shortform	Optional	Description
--input-dir	-i	NO	Input directory containing script PDFs.
--output	-o	NO	Output path for the CSV.
--page	-p	YES	Specify the page to conduct OMR on (defaults to first).
--rename	-r	YES	Flag for whether to rename documents based on their student numbers.
---save-level	-s	YES	Specify saveimglvl of OMRChecker (defaults as 0).

Related tools

Refer to this gist if you need a script to break up a pdf of multiple document scans (with same number of pages) into individual pdf doucments.