Manifold-Regularized Selectable Factor Extraction for Semi-supervised Image Classification

Xin Shi, Chao Zhang, Fangyun Wei, Hongyang Zhang and Yiyuan She

Abstract

In many vision analytics-based applications such as image classification, we confront explosive growth of high-dimensional data. Thus, many feature selection and extraction methods have been proposed to reduce the computational cost and avoid over-fitting. Recently, a novel selectable factor extraction (SFE) framework is proposed to simultaneously perform feature selection and extraction, and is theoretically and practically proved to be effective in handling high-dimensional data. The algorithm is also quite efficient and easy to implement. Although it is advantageous in several aspects, SFE is only designed for either supervised or unsupervised learning, and is not suitable when there are limited labeled samples and a large number of unlabeled samples, since the data distribution knowledge is likely to be poorly exploited. To tackle this problem, we propose a novel manifold regularized SFE (MRSFE) framework for semi-supervised image classification. In MRSFE, the local structures of the whole dataset are preserved, and the data distribution is well exploited. By integrating the label information, low rank property of the features and data distribution knowledge, the proposed MRSFE could select and extract reliable discriminative features when the labeled samples are scarce. An efficient and easy-to-implement algorithm is designed to find the solutions. Extensive experimental results on a real-world image dataset demonstrate the superiority of our method.

Session

Poster 2

Files

Extended Abstract (PDF, 101K)
Paper (PDF, 279K)

DOI

10.5244/C.29.132
https://dx.doi.org/10.5244/C.29.132

Citation

Bibtex

@inproceedings{BMVC2015_132,
	title={Manifold-Regularized Selectable Factor Extraction for Semi-supervised Image Classification},
	author={Xin Shi and Chao Zhang and  Fangyun Wei and Hongyang Zhang and Yiyuan She},
	year={2015},
	month={September},
	pages={132.1-132.11},
	articleno={132},
	numpages={11},
	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.29.132},
	isbn={1-901725-53-7},
	url={https://dx.doi.org/10.5244/C.29.132}
}