Optičko raspoznavanje oznaka na ispitnim obrascima

undergraduate thesis

• Title (english): Optical mark recognition on test forms
• Author: Dino Benedetti
• Mentor: Mate Puljiz (mentor)
• Committee member: Mate Puljiz (predsjednik povjerenstva)
• Committee member: Lana Horvat Dmitrović (član povjerenstva)
• Committee member: Ana Žgaljić Keko (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-09-11, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj je rad usmjeren na razvoj rješenja optičkog raspoznavanja oznaka (OMR) na ispitnim
obrascima. To podrazumijeva klasifikaciju zacrnjenih kružića i optičko prepoznavanje
znakova (OCR) za rukovanje ispravcima koji su pisani rukom. Objašnjeni su postupci
pretprocesiranja i poravnavanja slike za obradu, segmentacije interesnih područja te
detekcije relevantnih elemenata na slici. Za prepoznavanje rukom pisanih znamenki
korištena je neuronska mreža trenirana na MNIST skupu podataka. Rješenje se temelji na
Python programskom jeziku i koristi OpenCV za obradu slike i Keras za model neuronske
mreže. Prikazani su rezultati korištenja programskog rješenja te njihova evaluacija, a na
kraju je dan pregled budućeg rada i mogućih poboljšanja.

Abstract (english)

This paper focuses on the development of a solution for optical mark recognition (OMR)
on test forms also known as bubble sheets. These include classification of marked bubbles
and optical character recognition (OCR) to handle handwritten corrections. The paper
describes procedures for preprocessing and alignment of images, segmentation of regions
of interest (ROIs) and detection of relevant elements in the image. A neural network
trained on the MNIST data set was used to recognize handwritten digits. The solution is
based on the Python programming language and uses OpenCV for image processing and
Keras for the neural network model. The results of using the software solution and their
evaluation are presented, and at the end, an overview of future work and possible
improvements are given.

• Keywords: OMR; OCR; računalni vid; ispitni obrazac; optičko raspoznavanje; Python; OpenCV; poravnavanje; TensorFlow; Keras; neuronske mreže; segmentacija
• Keywords (english): OMR; OCR; computer vision; bubble sheet; answer sheet; optical recognition; Python; OpenCV; alignment; TensorFlow; Keras; neural networks; segmentation
• Language: croatian
• URN:NBN: urn:nbn:hr:168:720555
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 08:24:54