Diagnosing state-of-the-art object proposal methods

Hongyuan Zhu, Shijian Lu, Jianfei Cai and Guangqing Lee

Abstract

bject proposal has become a popular paradigm to replace exhaustive sliding window search in current top-performing methods in PASCAL VOC and ImageNet. Recently, Hosang et al. [17] conduct the first unified study of existing methods’ in terms of various image-level degradations. On the other hand, the vital question 'what objectlevel characteristics really affect existing methods’ performance?' is not yet answered. Inspired by Hoiem et al.’s work in categorical object detection, this paper conducts the first meta-analysis of various object-level characteristics’ impact on state-of-the-art object proposal methods. Specifically, we examine the effects of object size, aspect ratio, iconic view, color contrast, shape regularity and texture. We also analyse existing methods’ localization accuracy and latency for various PASCAL VOC object classes. Our study reveals the limitations of existing methods in terms of non-iconic view, small object size, low color contrast, shape regularity etc. Based on our observations, lessons are also learned and shared with respect to the selection of existing object proposal technologies as well as the design of the future ones

Session

Poster 1

Files

Extended Abstract (PDF, 98K)
Paper (PDF, 2M)

DOI

10.5244/C.29.11
https://dx.doi.org/10.5244/C.29.11

Citation

Bibtex

@inproceedings{BMVC2015_11,
	title={Diagnosing state-of-the-art object proposal methods},
	author={Hongyuan Zhu and Shijian Lu and Jianfei Cai and Guangqing Lee},
	year={2015},
	month={September},
	pages={11.1-11.12},
	articleno={11},
	numpages={12},
	booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
	publisher={BMVA Press},
	editor={Xianghua Xie, Mark W. Jones, and Gary K. L. Tam},
	doi={10.5244/C.29.11},
	isbn={1-901725-53-7},
	url={https://dx.doi.org/10.5244/C.29.11}
}