Prepoznavanje grafičkih znakova

undergraduate thesis

• Title (english): Recognition of graphical characters
• Author: Ivan Kraljević
• Mentor: Dragan Jevtić (mentor)
• Committee member: Dragan Jevtić (predsjednik povjerenstva)
• Committee member: Aleksandar Antonić (član povjerenstva)
• Committee member: Darko Štriga (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-09-20, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Tema ovog završnog rada bila je prepoznavanje grafičkih znakova sa slike korištenjem strojnog učenja. U svrhu strojnog izdvajanja korišten je KNN algoritam, a platforma korištena za obradu slike, u ovom radu je OpenCV.
KNN algoritam odluku o klasifikaciji odgađa sve do trenutka predočavanja upita. Točnost izdvajanja simbola sa slike varira u ovisnosti o samoj kvaliteti slike, o obradi klasifikacijskih podataka, kao i o pravilnom odabiru hiperparametra K. Izdvajanje simbola sa slike obrađeno je na primjeru prepoznavanja automobilskih tablica.
Ovaj rad sadrži opis svih postupaka potrebnih za pripremu slike za izdvajanje simbola, koji osim za ovaj specifični slučaj vrijede i za sve ostale primjere, također opisan je način rada KNN algoritma, sve njegove prednosti i mane, kao i postupak klasifikacije podatka za daljnji rad algoritma.

Abstract (english)

The subject of this thesis was the analysis of character recognition from image using machine learning. In the observed solution, KNN algorithm was used for character recognition. The platform used for image processing, in this thesis is OpenCV.
KNN algorithm postpones the decision of classification until the request comes. Accuracy of character recignition from image depends on quality of image, processing of the dataset for classification and on correctly choosing K parameter. Character recognition, in this thesis, was done on car license plates.
This thesis contains a description of all procedures that are needed for image preprocessing and for character recognition that are not only applicable in this specific case, but for almost every other case where a character or object is needed to be recognized. Furthermore, it contains an analysis of KNN algorithm with all his pros and cons, and procedures of classification of datasets that will later be used for recognition.

• Keywords: strojno učenje; izdvajanje simbola; metoda K najbližih susjeda; OpenCV
• Keywords (english): machine learning; character recognition; K nearest neighbour; OpenCV
• Language: croatian
• URN:NBN: urn:nbn:hr:168:565600
• 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: 2019-03-13 21:55:05