Representational Oriented Component Analysis (ROCA) for Face Recognition with One Sample Image per Training Class
Abstract
Subspace methods such as PCA, LDA, ICA have become a standard tool to perform visual learning and recognition. In this paper we propose Representational Oriented Component Analysis (ROCA), an extension of OCA, to perform face recognition when just one sample per training class is available. Several novelties are introduced in order to improve generalization and efficiency: • Combining several OCA classifiers based on different image representations of the unique training sample is shown to greatly improve the recognition performance. • To improve generalization and to account for small misregistration effect, a learned subspace is added to constrain the OCA solution, • A stable/efficient generalized eigenvector algorithm that solves the small size sample problem and avoids overfitting. Preliminary experiments in the FRGC Ver 1.0 dataset
Cite
Text
De la Torre et al. "Representational Oriented Component Analysis (ROCA) for Face Recognition with One Sample Image per Training Class." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2005. doi:10.1109/CVPR.2005.301Markdown
[De la Torre et al. "Representational Oriented Component Analysis (ROCA) for Face Recognition with One Sample Image per Training Class." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2005.](https://mlanthology.org/cvpr/2005/latorre2005cvpr-representational/) doi:10.1109/CVPR.2005.301BibTeX
@inproceedings{latorre2005cvpr-representational,
title = {{Representational Oriented Component Analysis (ROCA) for Face Recognition with One Sample Image per Training Class}},
author = {De la Torre, Fernando and Gross, Ralph and Baker, Simon and Kumar, B. V. K. Vijaya},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2005},
pages = {266-273},
doi = {10.1109/CVPR.2005.301},
url = {https://mlanthology.org/cvpr/2005/latorre2005cvpr-representational/}
}