Lokalizacija i očitavanja rukom pisanih bodova studenta

undergraduate thesis

• Title (english): Localization and Recognition of Handwritten Student Scores
• Author: Domagoj Pluščec
• Mentor: Marko Čupić (mentor)
• Committee member: Marko Čupić (predsjednik povjerenstva)
• Committee member: Miroslav Popović (član povjerenstva)
• Committee member: Mladen Karan (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Prilikom ispravljanja ispita koje piše veliki broj studenata ukazuje se potreba za sustavom za raspoznavanje rukom pisanih bodova studenata sa skeniranih slika.
U okviru ovog završnog rada opisan je i implementiran sustav za lokalizaciju i očitavanja rukom pisanih bodova studenata. Proces obrade obrasca za ocjenjivanje se sastojao od skeniranja obrasca, obrade slike, segmentacije znamenki bodova, klasifikacije znamenki te stvaranja izlazne datoteke.
Za obradu obrasca odabrani su algoritam minimalnog intenziteta za pretvaranje boje u nijanse sive te globalni algoritam fiksnog praga za binarizaciju slike. Segmentacija slike je provedena koristeći algoritam povezanih komponenti nakon kojeg je provedeno filtriranje suvišnih segmenata. Korišteni postupak segmentacije je postizao ukupnu stopu uspješne segmentacije od 78.51%. Klasifikacija znamenaka je provedena koristeći unaprijednu slojevitu neuronsku mrežu treniranu algoritmom propagacije pogreške unatrag. Korišteni klasifikator je postizao 82.41% uspješne stope klasifikacije na skupu primjera znamenki za testiranje.
Rezultat ukupne obrade obrazaca je iznosio 38% uspješne obrade ćelija od ukupno 1848 obrađenih ćelija.

Abstract (english)

When examining a large number of student exams, there is a need for an automatic system for localization and recognition of handwritten student scores.
In this bachelor's thesis a system for localization and recognition of handwritten student scores has been analysed and implemented. The analysis process consisted of form scanning, image processing, digit segmentation, digit classification and output file creation.
For processing the form, a minimal intensity algorithm for converting image to grayscale was used and a fixed global threshold algorithm for image binarization. Number segmentation was performed using the connected-component labeling algorithm after which the filtering of the redundant segments was performed. The segmentation process used achieved the overall rate of successful segmentation of 78.51%. Digit classification was performed using feedforward neural network trained with backpropagation algorithm. The classifier used achieved successful classification rate of 82.41% on test set. The total result of the form processing system was 38% of the total of 1848 processed form fields.

• Keywords: raspoznavanje rukom pisanih znamenaka; segmentacija znamenaka; klasifikacija znamenaka; neuronska mreža; propagacija pogreške unazad; lokalizacija teksta; obrada slike; algoritam označavanja povezanih komponenti
• Keywords (english): handwritten digit recognition; digit segmentation; digit classification; feedforward neural network; backpropagation; text localization; image processing; ocr; connected-component labeling algorithm
• Language: croatian
• URN:NBN: urn:nbn:hr:168:387599
• 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: 2019-03-14 00:03:13