Animal Recognition Using Different Deep Convolutional Neural Networks

Karsens, W. (2016) Animal Recognition Using Different Deep Convolutional Neural Networks. Bachelor's Thesis, Artificial Intelligence.

Preview	Text AI_BA_2016_wijnandkarsens.pdf - Published Version Download (973kB) | Preview
	Text Toestemming.pdf - Other Restricted to Backend only Download (572kB)

Abstract

This thesis describes the use of four Deep Convolutional Neural Network techniques for training and classifying various kinds of animals. To achieve this aim, several convolutional neural network architectures are used. The AlexNet, LeNet, GoogLeNet, and FlickrStyle architectures are used to train and evaluate the classification performance on three animal datasets: The Wild-Anim dataset and two other novel species-specific animal datasets: the RUG-Goats and RUG-Snakes datasets. These datasets are considered challenging due to the limited number of images per class and less discriminatory features between similar kinds of animal species. The experimental activities on the deep neural network architectures on these datasets were carried out on the Caffe deep learning framework. The results show that GoogLeNet outperforms all other Convolutional Neural Network (CNN) methods. Details of the experimental settings and the results are discussed in this paper.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Wiering, M. and Okafor, E.
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	15 Feb 2018 08:25
Last Modified:	02 May 2019 10:50
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/14589

Actions (login required)

View Item