Generating Discriminative Object Proposals via Submodular Ranking

Yangmuzi Zhang, Zhuolin Jiang, Xi Chen, Larry S. Davis; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2016, pp. 95-103

Abstract

A multi-scale greedy-based object proposal generation approach is presented. Our approach is built on top of a hierarchical segmentation. Object proposals are obtained by selecting a subset from the multi-scale segment pool via maximizing a submodular objective function, which consists of a weighted coverage term, a single-scale diversity term and a multi-scale reward term. The weighted coverage term forces the selected set of object proposals to be representative; single-scale diversity term encourages choosing segments from different exemplar clusters so that they will cover as many object patterns as possible, multi-scale reward term encourages the proposals to be discriminative and selected from multiple layers generated by the hierarchical image segmentation. The experimental results on the Berkeley Segmentation Dataset and PASCAL VOC2012 segmentation dataset demonstrate the accuracy and efficiency of our object proposal model.

Related Material

[pdf]
[bibtex]
@InProceedings{Zhang_2016_CVPR_Workshops,
author = {Zhang, Yangmuzi and Jiang, Zhuolin and Chen, Xi and Davis, Larry S.},
title = {Generating Discriminative Object Proposals via Submodular Ranking},
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2016}
}