Discrete Semi-supervised Multi-label Learning for Image Classification

Liang Xie, Lang He, Haohao Shu, Shengyuan Hu

Abstract: Multi-label image classification is a critical problem in semantic based image processing. Traditional semi-supervised multi-label learning methods usually learn classification functions in continuous label space. And the ignorance of discrete constraint of semantic labels impedes the classification performance. In this paper, we specifically consider the discrete constraint and propose Discrete Semi-supervised Multi-label Learning (DSML) for image classification. In DSML, we propose a semi-supervised framework with discrete constraint. Then we introduce anchor graph learning to improve the scalability, and derive an ADMM based alternating optimization process to solve the framework. Experimental results demonstrate the superiorly of DSML compared with several advanced semi-supervised methods.