Face Hallucination and Recognition Using Kernel Canonical Correlation Analysis
- 2.6k Downloads
Canonical correlation analysis (CCA) is a classical but powerful tool for image super-resolution tasks. Since CCA in essence is a linear projection learning method, it usually fails to uncover the nonlinear relationships between high-resolution (HR) and low-resolution (LR) facial image features. In order to solve this issue, we propose a new face hallucination and recognition algorithm based on kernel CCA, where the nonlinear correlation between HR and LR face features can be well depicted by implicit high-dimensional nonlinear mappings determined by specific kernels. First, our proposed method respectively extracts the principal component features from high-resolution and low-resolution facial images for computational efficiency and noise removal. Then, it makes use of kernel CCA to learn the nonlinear consistency of HR and LR facial features. The proposed approach is compared with existing face hallucination algorithms. A number of experimental results on LR face recognition have demonstrated the effectiveness and robustness of our proposed method.
KeywordsFace hallucination Kernel CCA Face recognition
This work is supported by National Natural Science Foundation of China under Grant No. 61402203. In addition, it is also supported in part by the National Natural Science Foundation of China under Grant Nos. 61472344, 61611540347, the Natural Science Foundation of Jiangsu Province of China under Grant Nos. BK20161338, BK20170513, and sponsored by Excellent Young Backbone Teacher Project.
- 1.Baker, S., Kanade, T.: Hallucinating faces. In: 4th IEEE International Conference on Automatic Face and Gesture Recognition, pp. 83–88. IEEE, Grenoble (2000)Google Scholar
- 2.Chang, H., Yeung, D.Y., Xiong Y.: Super-resolution through neighbor embedding. In: CVPR, pp. 275–282. IEEE, Washington (2004)Google Scholar
- 3.Glasner, D., Bagon, S., Irani, M.: Super-resolution from a single image. In: ICCV, pp. 349–356. IEEE, Kyoto (2009)Google Scholar
- 4.Lu, X., Yuan, Y., Yan, P.: Image super-resolution via double sparsity regularized manifold learning. IEEE T-CSVT 23(12), 2022–2033 (2013)Google Scholar
- 8.He, L., Qi, H., Zaretzki, R.: Beta process joint dictionary learning for coupled feature spaces with application to single image super-resolution. In: CVPR, pp. 345–352. IEEE, Portland (2013)Google Scholar
- 9.Zeyde, R., Elad, M., Protter, M.: On single image scale-up using sparse-representations. In: Boissonnat, J.-D., Chenin, P., Cohen, A., Gout, C., Lyche, T., Mazure, L., Schumaker, L. (eds.) Curves and Surfaces 2010. LNCS, vol. 6920, pp. 711–730. Springer, Heidelberg (2012). doi: 10.1007/978-3-642-27413-8_47 CrossRefGoogle Scholar
- 13.Kim, J., Lee, J.K., Lee K.M.: Accurate image super-resolution using very deep convolutional networks. In: CVPR, pp. 1646–1654. IEEE, Las Vegas (2016)Google Scholar
- 14.Shi, W., Caballero, J., Huszar, F., Totz, J., Aitken, A.P., Bishop, R., Rueckert, D., Wang, Z.: Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In: CVPR, pp. 1874–1883. IEEE, Las Vegas (2016)Google Scholar