Klasifikacija slika konvolucijskim neuronskim mrežama

undergraduate thesis

• Title (english): Image Classification with Deep Neural Networks
• Author: Ivana Babić
• Mentor: Siniša Šegvić (mentor)
• Committee member: Siniša Šegvić (predsjednik povjerenstva)
• Committee member: Mladen Karan (član povjerenstva)
• Committee member: Ivan Krešo (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-09-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Tema završnog rada jest klasifikacija slika konvolucijskim neuronskim mrežama.
Kroz zadnjih nekoliko godina korištenje dubokih neuronskih mreža dovelo je do postizanja jako dobrih rezultata u rješavanju zadataka iz područja računalnog vida, a za postupke vizualne klasifikacije najboljima su se pokazale konvolucijske neuronske mreže.
Rad obuhvaća teoretsku podlogu vezanu za konvolucijske mreže, od same definicije mreže, značaja koji danas ima u zadatcima veznim za klasifikaciju slika, tipove i prostorni raspored slojeva i neurona te opis mreža koje su izrađene u sklopu ovog rada.
Također je dan osvrt na TensorFlow, Googleovu biblioteku koja je korištena za programsku izvedbu zadataka.

Abstract (english)

The topic of this paper is image classification using deep neural networks.
During last several years, the use of deep neural networks has led to breakthrough results on a variety of pattern recognition problems, such as computer vision. Convolutional neural network is a special kind of neural network most used in visual classification problems.
This paper gives a theoretical background referring to ConvNets, the significance they have in achieving state-of-the-art results in modern image classification problems, spatial arrangement, layer patterns and type of neurons used for making the network that was a part of this assignment.
A brief review of TensorFlow is also given, due to the fact that TensorFlow is the library used in making of this project.

• Keywords: duboke neuronske mreže; konvolucijska neuronska mreža; klasifikacija slika; ConvNets; CIFAR-10; ReLU; TensorFlow; Python
• Keywords (english): Deep learning; convolutional neural network; image classification; ConvNets; CIFAR-10; ReLU; TensorFlow; Python
• Language: croatian
• URN:NBN: urn:nbn:hr:168:465056
• 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 18:41:43