As multiple-choice assessments are constantly being used to acquire assessment data on a given subject, there is a need for accessible and low-cost methods of reading these assessments for correctors. The answer sheet correction process is done manually or with the use of special optical readers for each type of form, hence, there is the need for a system that corrects this forms through the use of a library using a personal scanner. In this way, this work proposed the development of a library for the optical mark recognition (OMR) in forms by the use of digital image processing techniques. The library was developed in Python language, being multiplatform and capable of to detecting templates and alternatives in different formats of answer sheets and going thorough them, it is capable of detecting the number of pixels for each question, determining which are the marks and validating them, and detect the position of each mark. The results acquired through objects such as QR-Codes or barcodes, questions and marks are sent to a JSON file.
- Opencv --version 4.1.2.30
- pip --version 19.0.3
- imutils --version 0.5.3
- numpy --version 1.18.1
- pyzbar --version 0.1.8
- scipy --version 1.4.1
- Python 3.7
- Correct angle
- OMR
- file .json
- tests unity
