prikaz prve stranice dokumenta Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja
  Download
PDF 2.21 MB
undergraduate thesis
Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja
2019. urn:nbn:hr:200:141378
Mihelčić, Dario
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Software Engineering
Chair of Programming Languages and Systems

Cite this document

Mihelčić, D. (2019). Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja (Undergraduate thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:141378

Mihelčić, Dario. "Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2019. https://urn.nsk.hr/urn:nbn:hr:200:141378

Mihelčić, Dario. "Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2019. https://urn.nsk.hr/urn:nbn:hr:200:141378

Mihelčić, D. (2019). 'Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja', Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 22 November 2024, https://urn.nsk.hr/urn:nbn:hr:200:141378

Mihelčić D. Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja [Undergraduate thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2019 [cited 2024 November 22] Available at: https://urn.nsk.hr/urn:nbn:hr:200:141378

D. Mihelčić, "Android mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja", Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2019. Available at: https://urn.nsk.hr/urn:nbn:hr:200:141378

Please login to the repository to save this object to your list.
Metadata
TitleAndroid mobilna aplikacija za prepoznavanje pasmina pasa postupcima strojnog učenja
Title (english)Android mobile application for dog breed recognition based on machine learning
AuthorDario Mihelčić
MentorGoran Martinović (mentor)
Committee memberGoran Martinović (predsjednik povjerenstva)
Committee memberZdravko Krpić (član povjerenstva)
Committee memberČaslav Livada (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Software Engineering)
(Chair of Programming Languages and Systems)
Osijek
Defense date and country2019-10-03, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Program Engineering
Abstract
Cilj ovog rada bio je izraditi model i mobilnu aplikaciju koja može prepoznati pasmine pasa sa slike. U teorijskom dijelu rada objašnjeni su postupci strojnog učenja koji se mogu koristiti za prepoznavanje pasmina sa slika. Nakon danih teorijskih opisa postupaka te istraživanjem primjena postupaka pronađene su neuronske mreže kao najbolje rješenje. Nakon toga, dan je opis alata i algoritama koji se koriste u ostvarenju odabranog postupka. Odabrana je metoda prenesenog učenja za stvaranje modela za prepoznavanje. U programskom dijelu rada izrađene su funkcije za jednostavnu izgradnju modela za prepoznavanje odabranih klasa. Izrađena su četiri modela za klasificiranje 133 psećih pasmina od kojih je testiranjem izabran najbolji model. Izrađena je Android aplikacija u kojoj je model implementiran. Uz mogućnost prepoznavanja pasmina pasa, aplikacija daje korisniku i opise svih pasmina. Rezultati testiranja aplikacije pokazali su određene probleme koji mogu nastati kod klasifikacije objekata sa slika kao što je raznolikost karakteristika članova jedne klase ili zajedničke karakteristike više klasa.
Abstract (english)
The goal of this paper was to create a model and a mobile application that can recognize dog breeds in pictures. Methods of machine learning which can be used for recognizing dog breeds from pictures were explained in the theoretical part of the paper. After the theoretical description of the methods was given and usages of those methods were explored, neural networks were found to be the best solution to the problem. After that, a description of tools and algorithms used to create the solution to the problem were given. Transfer learning was chosen for creating the recognition model. In the programming part of the paper, functions for easy creation of recognition models capable of recognizing the chosen classes were made. Four models capable of recognizing 133 dog breeds were made from which the best one was chosen through testing. Apart from recognizing dog breeds, the application gives the user description of all dog breeds. While testing the applications, some problems that can occur in object classification were found, such as differences in characteristics of members of the same class or characteristics shared among different classes.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:141378
Study programmeTitle: Undergraduate University Study Programme in Computer Engineering
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 resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2019-12-20 15:36:05