Weakly-Supervised Discovery of Visual Pattern Configurations

TitleWeakly-Supervised Discovery of Visual Pattern Configurations
Publication TypeConference Paper
Year of Publication2014
AuthorsSong, H. Oh, Lee Y. Jae, Jegelka S., & Darrell T.
Other Numbers3699
Abstract

The prominence of weakly labeled data gives rise to a growing demand for object detection methods that can cope with minimal supervision. We propose anapproach that automatically identifies discriminative configurations of visual patterns that are characteristic of a given object class. We formulate the problem as aconstrained submodular optimization problem and demonstrate the benefits of thediscovered configurations in remedying mislocalizations and finding informativepositive and negative training examples. Together, these lead to state-of-the-artweakly-supervised detection results on the challenging PASCAL VOC dataset.

URLhttp://www.icsi.berkeley.edu/pubs/vision/weeklysupervised14.pdf
Bibliographic Notes

Proceedings of the 28th Annual Conference on Neural Information Processing Systems (NIPS 2014), Montreal, Canada

Abbreviated Authors

H. O. Song, Y. J. Lee, S. Jegelka, and T. Darrell

ICSI Research Group

Vision

ICSI Publication Type

Article in conference proceedings