Conference paper

AHMAD Touqeer, CAMPR Pavel, ČADÍK Martin and BEBIS George. Comparison of Semantic Segmentation Approaches for Horizon/Sky Line Detection. In: Proceedings of the International Joint Conference on Neural Networks (IJCNN). Anchorage: Institute of Electrical and Electronics Engineers, 2017, pp. 1-8. ISBN 978-1-4799-1961-1. Available from: http://cadik.posvete.cz/papers/ahmad17comparison.pdf
Publication language:english
Original title:Comparison of Semantic Segmentation Approaches for Horizon/Sky Line Detection
Title (cs):Porovnání metod sémantické segmentace pro detekci horizontu
Pages:1-8
Proceedings:Proceedings of the International Joint Conference on Neural Networks (IJCNN)
Conference:The 2017 International Joint Conference on Neural Networks
Place:Anchorage, US
Year:2017
URL:http://cadik.posvete.cz/papers/ahmad17comparison.pdf
ISBN:978-1-4799-1961-1
Publisher:Institute of Electrical and Electronics Engineers
Files: 
+Type Name Title Size Last modified
iconahmad17comparison.pdf7,46 MB2017-03-14 16:30:42
^ Select all
With selected:
Keywords
horizon detection, skyline detection, geo-localization, segmentation methods, semantic segmentation, Fully Convolutional Networks
Annotation
Horizon or skyline detection plays a vital role towards mountainous visual geo-localization, however most of the recently proposed visual geo-localization approaches rely on user-in-the-loop skyline detection methods. Detecting such a segmenting boundary fully autonomously would definitely be a step forward for these localization approaches. This paper provides a quantitative comparison of four such methods for autonomous horizon/sky line detection on an extensive data set. Specifically, we provide the comparison between four recently proposed segmentation methods; one explicitly targeting the problem of horizon detection, second focused on visual geo-localization but relying on accurate detection of skyline and other two proposed for general semantic segmentation -- Fully Convolutional Networks (FCN) and SegNet. Each of the first two methods is trained on a common training set comprised of about 200 images while models for the third and fourth method are fine tuned for sky segmentation problem through transfer learning using the same data set. Each of the method is tested on an extensive test set (about 3K images) covering various challenging geographical, weather, illumination and seasonal conditions. We report average accuracy and average absolute pixel error for each of the presented formulation.

BibTeX:
@INPROCEEDINGS{
   author = {Touqeer Ahmad and Pavel Campr and Martin {\v{C}}ad{\'{i}}k
	and George Bebis},
   title = {Comparison of Semantic Segmentation Approaches for
	Horizon/Sky Line Detection},
   pages = {1--8},
   booktitle = {Proceedings of the International Joint Conference on Neural
	Networks (IJCNN)},
   year = {2017},
   location = {Anchorage, US},
   publisher = {Institute of Electrical and Electronics Engineers},
   ISBN = {978-1-4799-1961-1},
   language = {english},
   url = {http://www.fit.vutbr.cz/research/view_pub.php.en.iso-8859-2?id=11352}
}

Your IPv4 address: 54.81.45.122
Switch to IPv6 connection

DNSSEC [dnssec]