Common Subspace Based Low-Rank and Joint Sparse Representation for Multi-view Face Recognition
Multi-view face data are very common in real-world application, since different viewpoints and various types of sensors attempt to better represent face data. However, these data have large pose variation, which dramatically degrades the performance of multi-view face recognition. To address this, we propose a common subspace based low-rank and joint sparse representation (CSLRJSR) method, which provides a framework encompassing divergence mitigation and feature fusion. In CSLRJSR method, common subspace is learnt to bridge the view, then low-rank and joint sparse representation are exploited to learn and then fuse the discriminative features. Experiments on multi-view face dataset demonstrate that CSLRJSR outperforms the state-of-the-art methods both in two-view and multi-view situations.
KeywordsMulti-view face recognition Common subspace Low-rank representation Joint sparse
The work was supported by National Natural Science Foundation of China (No. 61803075). We thank the anonymous reviewers for their comments and suggestions which make the paper much improved.
- 2.Cao, L., Luo, J., Liang F., Huang, T.: Heterogeneous feature machines for visual recognition. In: IEEE International Conference on Computer Vision (2009)Google Scholar
- 5.Yang, J., Wright, J., Huang, T., Ma, Y.: Image super-resolution as sparse representation of raw image patches. In: 2008 Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1–8 (2008)Google Scholar
- 6.Wang, S., Zhang, L., Liang, Y., Pan, Q.: Semi-coupled dictionary learning with applications to image super-resolution and photo-sketch synthesis. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2216–2223 (2012)Google Scholar
- 7.Huang D., Wang, Y.: Coupled dictionary and feature space learning with applications to cross-domain image synthesis and recognition. In: IEEE International Conference on Computer Vision, pp. 2496–2503 (2013)Google Scholar
- 12.Ding, Z., Fu, Y.: Low-rank common subspace for multi-view learning. In: Proceedings of the IEEE International Conference on Data Mining, pp. 110–119 (2014)Google Scholar
- 16.Zhang, H., Patel, V., Chellappa, R.: Robust multimodal recognition via multitask multivariate low-rank representations. In: 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG), pp. 1–8 (2015)Google Scholar
- 20.Li, S., Fu, Y.: Robust subspace discovery through supervised low-rank constraints. In: Proceedings of SIAM International Conference on Data Mining, pp. 163–171 (2014)Google Scholar
- 24.Schölkopf, B., Platt, J., Hofmann, T.: Efficient sparse coding algorithms. In: Advances in Neural Information Processing Systems, pp. 801–808 (2007)Google Scholar
- 25.Sim, T., Baker, S., Bsat, M.: The CMU pose, illumination, and expression (PIE) database of human faces. In: Proceedings of Fifth IEEE International Conference on Automatic Face Gesture Recognition, pp. 53–58 (2002)Google Scholar