undergraduate thesis
Neural networks in OpenCV library

Bože Eugen Marković (2016)
Metadata
TitleNeuronske mreže u OpenCV biblioteci
AuthorBože Eugen Marković
Mentor(s)Damir Filko (thesis advisor)
Abstract
Cilj ovog završnog rada bio je prikazati teorijsku podlogu umjetnih neuronskih mreža i njihovu primjenu u C++ programskom jeziku sa OpenCV bibliotekom. Obrađeni problem je prepoznavanje ljudskog rukopisa, točnije znamenaka. Potrebno je stvoriti neuronsku mrežu i definirati njene parametre što se postiže s klasama i funkcijama iz implementirane OpenCV biblioteke. Pokreće se proces učenja neuronske mreže sa što većim brojem uzoraka. Kad je proces treniranja gotov sustavu se predaju testni uzorci kako bi se utvrdila uspješnost procesa učenja. U simulaciji je mreža trenirana sa 1000 uzoraka, a testnih primjera predano joj je 100. Od 100 uzoraka prepoznato ih je 83 što bi značilo da je uspješnost na relativno visokoj razini.
Keywordsneural network OpenCV library C++ language handwriting recognition
Parallel title (English)Neural networks in OpenCV library
Committee MembersDamir Filko (committee chairperson)
Hrvoje Glavaš (committee member)
Damir Blažević (committee member)
GranterSveučilište Josipa Jurja Strossmayera u Osijeku
Fakultet elektrotehnike, računarstva i informacijskih tehnologija Osijek
Lower level organizational unitsZavod za računalno inženjerstvo i automatiku
Katedra za automatiku i robotiku
PlaceOsijek
StateCroatia
Scientific field, discipline, subdisciplineTECHNICAL SCIENCES
Computing
Process Computing
Study programme typeuniversity
Study levelundergraduate
Study programmeUndergraduate University Study Programme in Computer Engineering
Academic title abbreviationuniv.bacc.ing.comp.
Genreundergraduate thesis
Language Croatian
Defense date2016-09-28
Parallel abstract (English)
The aim of this bachelor thesis was to show the theoretical basis of artificial neural networks and their use in C ++ programming language with OpenCV library. Problem that was processed is human handwriting recognition, digits. It is necessary to create a neural network and define its parameters using classes and functions from the OpenCV library. The process of learning neural network is initiated with training samples. When the training process is finished, test samples are used in order to determine the success of the learning process. In the simulation, the network is trained with a 1000 samples and given a 100 test examples. Of the 100 samples, 83 were identified, which would mean that the success rate is relatively high.
Parallel keywords (Croatian)neuronska mreža OpenCV biblioteka C++ jezik prepoznavanje rukopisa
Resource typetext
Access conditionOpen access
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:200:991934
CommitterAnka Ovničević