Contour Box: Rejecting Object Proposals Without Explicit Closed Contours

Cewu Lu, Shu Liu, Jiaya Jia, Chi-Keung Tang

ICCV 2015

doi:10.1109/ICCV.2015.234 /iccv/2015/lu2015iccv-contour/

Abstract

Closed contour is an important objectness indicator. We propose a new measure subject to the completeness and tightness constraints, where the optimized closed contour should be tightly bounded within an object proposal. The closed contour measure is defined using closed path integral, and we solve the optimization problem efficiently in polar coordinate system with a global optimum guaranteed. Extensive experiments show that our method can reject a large number of false proposals, and achieve over 6% improvement in object recall at the challenging overlap threshold 0.8 on the VOC 2007 test dataset.

PDF ICCV Semantic Scholar

Cite

Text

Lu et al. "Contour Box: Rejecting Object Proposals Without Explicit Closed Contours." International Conference on Computer Vision, 2015. doi:10.1109/ICCV.2015.234

Markdown

[Lu et al. "Contour Box: Rejecting Object Proposals Without Explicit Closed Contours." International Conference on Computer Vision, 2015.](https://mlanthology.org/iccv/2015/lu2015iccv-contour/) doi:10.1109/ICCV.2015.234

BibTeX

@inproceedings{lu2015iccv-contour,
  title     = {{Contour Box: Rejecting Object Proposals Without Explicit Closed Contours}},
  author    = {Lu, Cewu and Liu, Shu and Jia, Jiaya and Tang, Chi-Keung},
  booktitle = {International Conference on Computer Vision},
  year      = {2015},
  doi       = {10.1109/ICCV.2015.234},
  url       = {https://mlanthology.org/iccv/2015/lu2015iccv-contour/}
}