diplomsko delo
Bojan Jan Javornik (Author), Vili Podgorelec (Mentor)

Abstract

Cilj diplomske naloge je preizkusiti klasifikacijo slik s pomočjo globokih nevronskih mrež. Odločili smo se za uporabo konvolucijskih nevronskih mrež, saj so najbolj razširjene na področju klasifikacije slik, učne in testno množico pa smo pripravili sami. Modele mreže smo gradili in učili v programskem jeziku Python s pomočjo knjižnice Keras. Opišemo kako smo slike pretvorili v vhode konvolucijske nevronske mreže in kakšne modele smo zgradili. Primernost modelov smo ugotavljali z navzkrižno validacijo, nato pa smo jih preizkusili še na testni množici. Vse rezultate tudi predstavimo in povemo ugotovitve.

Keywords

globoko učenje;konvolucijske nevronske mreže;klasifikacija slik;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [B. J. Javornik]
UDC: 004.8:004.932(043.2)
COBISS: 22809878 Link will open in a new window
Views: 792
Downloads: 69
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: The preparation of photos' dataset and its classification using deep neural networks
Secondary abstract: The main goal of this thesis is the examination of image classification using deep neural networks. We have decided to build our models with convolutional neural networks, because they are the most popular algorithm for image classification. We also prepared our own train and test datasets. Models were built and trained in programming language Python with the help of Keras library. We describe how we transformed datasets into inputs of convolutional neural network and what models have we built. We tested models' adequacy using cross validation then we tested their performance on test dataset. We present our results and introduce our findings.
Secondary keywords: deep learning;convolutional neural networks;image classification;
Type (COBISS): Bachelor thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Informatika in tehnologije komuniciranja
Pages: VII, 37 str.
ID: 11215930