Foreground Segmentation via Dynamic Programming

Bing LUO; Chao HUANG; Lei MA; Wei LI; Qingbo WU

doi:10.1587/transinf.2014EDL8078

Foreground Segmentation via Dynamic Programming

Bing LUO, Chao HUANG, Lei MA, Wei LI, Qingbo WU

Full Text Views

0

Share
Cite this

Summary :

This paper proposes a novel method to segment the object of a specific class based on a rough detection window (such as Deformable Part Model (DPM) in this paper), which is robust to the positions of the bounding boxes. In our method, the DPM is first used to generate the root and part windows of the object. Then a set of object part candidates are generated by randomly sampling windows around the root window. Furthermore, an undirected graph (the minimum spanning tree) is constructed to describe the spatial relationships between the part windows. Finally, the object is segmented by grouping the part proposals on the undirected graph, which is formulated as an energy function minimization problem. A novel energy function consisting of the data term and the smoothness term is designed to characterize the combination of the part proposals, which is globally minimized by the dynamic programming on a tree. Our experimental results on challenging dataset demonstrate the effectiveness of the proposed method.

Publication: IEICE TRANSACTIONS on Information Vol.E97-D No.10 pp.2818-2822

Publication Date: 2014/10/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2014EDL8078

Type of Manuscript: LETTER

Category: Image Recognition, Computer Vision

Authors

Bing LUO
  University of Electronic Science and Technology of China
Chao HUANG
  University of Electronic Science and Technology of China
Lei MA
  University of Electronic Science and Technology of China
Wei LI
  University of Electronic Science and Technology of China
Qingbo WU
  University of Electronic Science and Technology of China

Keyword

object segmentation, undirected graph, dynamic programming

Cite this

Copy

Bing LUO, Chao HUANG, Lei MA, Wei LI, Qingbo WU, "Foreground Segmentation via Dynamic Programming" in IEICE TRANSACTIONS on Information, vol. E97-D, no. 10, pp. 2818-2822, October 2014, doi: 10.1587/transinf.2014EDL8078.
Abstract: This paper proposes a novel method to segment the object of a specific class based on a rough detection window (such as Deformable Part Model (DPM) in this paper), which is robust to the positions of the bounding boxes. In our method, the DPM is first used to generate the root and part windows of the object. Then a set of object part candidates are generated by randomly sampling windows around the root window. Furthermore, an undirected graph (the minimum spanning tree) is constructed to describe the spatial relationships between the part windows. Finally, the object is segmented by grouping the part proposals on the undirected graph, which is formulated as an energy function minimization problem. A novel energy function consisting of the data term and the smoothness term is designed to characterize the combination of the part proposals, which is globally minimized by the dynamic programming on a tree. Our experimental results on challenging dataset demonstrate the effectiveness of the proposed method.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2014EDL8078/_p

Copy

@ARTICLE{e97-d_10_2818,
author={Bing LUO, Chao HUANG, Lei MA, Wei LI, Qingbo WU, },
journal={IEICE TRANSACTIONS on Information},
title={Foreground Segmentation via Dynamic Programming},
year={2014},
volume={E97-D},
number={10},
pages={2818-2822},
abstract={This paper proposes a novel method to segment the object of a specific class based on a rough detection window (such as Deformable Part Model (DPM) in this paper), which is robust to the positions of the bounding boxes. In our method, the DPM is first used to generate the root and part windows of the object. Then a set of object part candidates are generated by randomly sampling windows around the root window. Furthermore, an undirected graph (the minimum spanning tree) is constructed to describe the spatial relationships between the part windows. Finally, the object is segmented by grouping the part proposals on the undirected graph, which is formulated as an energy function minimization problem. A novel energy function consisting of the data term and the smoothness term is designed to characterize the combination of the part proposals, which is globally minimized by the dynamic programming on a tree. Our experimental results on challenging dataset demonstrate the effectiveness of the proposed method.},
keywords={},
doi={10.1587/transinf.2014EDL8078},
ISSN={1745-1361},
month={October},}

Copy

TY - JOUR
TI - Foreground Segmentation via Dynamic Programming
T2 - IEICE TRANSACTIONS on Information
SP - 2818
EP - 2822
AU - Bing LUO
AU - Chao HUANG
AU - Lei MA
AU - Wei LI
AU - Qingbo WU
PY - 2014
DO - 10.1587/transinf.2014EDL8078
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E97-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2014
AB - This paper proposes a novel method to segment the object of a specific class based on a rough detection window (such as Deformable Part Model (DPM) in this paper), which is robust to the positions of the bounding boxes. In our method, the DPM is first used to generate the root and part windows of the object. Then a set of object part candidates are generated by randomly sampling windows around the root window. Furthermore, an undirected graph (the minimum spanning tree) is constructed to describe the spatial relationships between the part windows. Finally, the object is segmented by grouping the part proposals on the undirected graph, which is formulated as an energy function minimization problem. A novel energy function consisting of the data term and the smoothness term is designed to characterize the combination of the part proposals, which is globally minimized by the dynamic programming on a tree. Our experimental results on challenging dataset demonstrate the effectiveness of the proposed method.
ER -