Sustav za automatsko očitavanje rukom pisanih ispravaka odgovora na ispitnom obrascu

master's thesis

• Title (english): System for Automated Recognition of Handwritten Corrections for Multiple Choice Answer Sheets
• Author: Matea Cerovac
• Mentor: Zoran Kalafatić (mentor)
• Committee member: Zoran Kalafatić (predsjednik povjerenstva)
• Committee member: Tomislav Hrkać (član povjerenstva)
• Committee member: Alan Jović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-18, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Na pojedinim kolegijima Fakulteta elektrotehnike i računarstva koriste se ispitni
obrasci s mogućnošću zacrnjivanja odgovora. S obzirom da studenti ponekad pogriješe,
postoji mogućnost ispravka odgovora upisivanjem znaka željenog odgovora u
predviđeno polje. Takvi ispravci otežavaju proces automatiziranog ocjenjivanja, jer se
od ispravljača ispita koji koristi postojeći sustav traži očitavanje i ručni unos popravljenog
odgovora. Kako bi se smanjila potreba za takvim postupkom, treba osmisliti
i testirati metodu automatiziranog očitavanja ispravaka u obrascu, u vidu klasifikacije
konvolucijskom neuronskom mrežom. Za testiranje spomenute metode klasifikacije,
pripremljen je skup podataka za učenje i testiranje sustava te su analizirani dobiveni
rezultati.

Abstract (english)

For some courses at Faculty of Electrical Engineering and Computing there exist
multiple-choice exam forms that are annotated by students. The process of annotation
by the students is sometimes prone to errors, so the students have possibility to alter
their answer by writing correction to a predefined area in the form. Those corrections
complicate the process of automated grading, because the system needs help of an operator
for manual recognition and input of the corrections added on the form. In order to
simplify the process of the manual input of the corrections, a method for recognition
for the corrections is given in form of a classification by the use of a convolutional
neural network. In order to test the mentioned method, a training and testing dataset is
manually prepared and the results are analyzed.

• Keywords: Raspoznavanje rukopisa; konvolucijske neuronske mreže; ispitni obrasci
• Keywords (english): Handwriting recognition; convolutional neural networks; multiple-choice exam forms
• Language: croatian
• URN:NBN: urn:nbn:hr:168:282968
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 19:34:21