Weakly Supervised Object Localization with Large Fisher Vectors
Josip Krapac, Siniša Šegvić
2015
Abstract
We propose a novel method for learning object localization models in a weakly supervised manner, by employing images annotated with object class labels but not with object locations. Given an image, the learned model predicts both the presence of the object class in the image and the bounding box that determines the object location. The main ingredients of our method are a large Fisher vector representation and a sparse classification model enabling efficient evaluation of patch scores. The method is able to reliably detect very small objects with some intra-class variation in reasonable time. Experimental validation has been performed on a public dataset and we report localization performance comparable to strongly supervised approaches.
DownloadPaper Citation
in Harvard Style
Krapac J. and Šegvić S. (2015). Weakly Supervised Object Localization with Large Fisher Vectors . In Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP, (VISIGRAPP 2015) ISBN 978-989-758-090-1, pages 44-53. DOI: 10.5220/0005294900440053
in Bibtex Style
@conference{visapp15,
author={Josip Krapac and Siniša Šegvić},
title={Weakly Supervised Object Localization with Large Fisher Vectors},
booktitle={Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP, (VISIGRAPP 2015)},
year={2015},
pages={44-53},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005294900440053},
isbn={978-989-758-090-1},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP, (VISIGRAPP 2015)
TI - Weakly Supervised Object Localization with Large Fisher Vectors
SN - 978-989-758-090-1
AU - Krapac J.
AU - Šegvić S.
PY - 2015
SP - 44
EP - 53
DO - 10.5220/0005294900440053