Attention Networks for Weakly Supervised Object Localization

Eu Wern Teh, Mrigank Rochan and Yang Wang

Abstract

We consider the problem of weakly supervised learning for object localization. Given a collection of images with image-level annotation indicating the presence/absence of an object, our goal is to localize the object in each image. In this paper, we propose an deep network architecture called attention network for this problem. Given a set candidate regions in each image, the attention network first computes an attention score on each candidate region. Then these candidate regions are combined together with their attention scores to form a whole-image feature vector. This feature vector is used for classifying the image. The object localization is implicitly achieved via the attention scores on candidate regions. We demonstrate the our approach achieves performance comparable to or better than other state-of-the-art methods on several benchmark datasets.

Session

Posters 1

Files

Extended Abstract (PDF, 187K)
Paper (PDF, 563K)

DOI

10.5244/C.30.52
https://dx.doi.org/10.5244/C.30.52

Citation

Bibtex

        @inproceedings{BMVC2016_52,
        	title={Attention Networks for Weakly Supervised Object Localization},
        	author={Eu Wern Teh, Mrigank Rochan and Yang Wang},
        	year={2016},
        	month={September},
        	pages={52.1-52.11},
        	articleno={52},
        	numpages={11},
        	booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
        	publisher={BMVA Press},
        	editor={Richard C. Wilson, Edwin R. Hancock and William A. P. Smith},
        	doi={10.5244/C.30.52},
        	isbn={1-901725-59-6},
        	url={https://dx.doi.org/10.5244/C.30.52}
        }