BIROn - Birkbeck Institutional Research Online

    Shared feature extraction for semi-supervised image classification

    Luo, Y. and Tao, D. and Geng, B. and Xu, C. and Maybank, Stephen J. (2011) Shared feature extraction for semi-supervised image classification. In: Proceedings of the 19th ACM international conference on Multimedia: MM '11, 28 Nov - 01 Dec 2011, Scottsdale, U.S..

    Full text not available from this repository.

    Abstract

    Multi-task learning (MTL) plays an important role in image analysis applications, e.g. image classification, face recognition and image annotation. That is because MTL can estimate the latent shared subspace to represent the common features given a set of images from different tasks. However, the geometry of the data probability distribution is always supported on an intrinsic image sub-manifold that is embedded in a high dimensional Euclidean space. Therefore, it is improper to directly apply MTL to multiclass image classification. In this paper, we propose a manifold regularized MTL (MRMTL) algorithm to discover the latent shared subspace by treating the high-dimensional image space as a sub-manifold embedded in an ambient space. We conduct experiments on the PASCAL VOC'07 dataset with 20 classes and the MIR dataset with 38 classes by comparing MRMTL with conventional MTL and several representative image classification algorithms. The results suggest that MRMTL can properly extract the common features for image representation and thus improve the generalization performance of the image classification models.

    Metadata

    Item Type: Conference or Workshop Item (Paper)
    School: Birkbeck Schools and Departments > School of Business, Economics & Informatics > Computer Science and Information Systems
    Depositing User: Administrator
    Date Deposited: 06 Nov 2012 11:18
    Last Modified: 17 Apr 2013 12:26
    URI: http://eprints.bbk.ac.uk/id/eprint/5567

    Statistics

    Downloads
    Activity Overview
    0Downloads
    65Hits

    Additional statistics are available via IRStats2.

    Archive Staff Only (login required)

    Edit/View Item Edit/View Item