/S3RC

Primary LanguageMatlab

Codes for Semi-Supervised Sparse Representation based Classificatin (S3RC), Version 1.0

Please refer to our following paper for algorithm details:

Yuan Gao, Jiayi Ma, and Alan L. Yuille, Semi-Supervised Sparse Representation Based Classification for Face Recognition with Insufficient Labeled Samples, IEEE Transactions on Image Processing, 2017. In Press.

Usage:

  • run S3RC_single_labeled_sample.m to start the demo for Face Recognition with insufficient labeled samples;
  • run S3RC_insufficient_labeled_samples.m to start the demo for Face Recognition with single labeled sample per person.
  • The demo codes needs L1_homotopy_v2.0 toobox to solve the L1 minimization problem (already included), which can be acquired from http://www.ece.ucr.edu/~sasif/homotopy/.

Features:

  • This release implements the vanilla version of our S3RC algorithm, i.e., our main contribution on the gallery dictionary learning with the basic ESRC variation dictionary.
  • In order to combine more advanced variation dictionary, e.g., S3RC-SVDL, S3RC-RADL, or your own variation dictionary learning, just replace the variation dictionary V in S3RC_single_labeled_sample.m or S3RC_insufficient_labeled_samples.m.

Contacts

For questions about the code or the paper, feel free to contact me by Ethan.Y.Gao@gmail.com.

Bibtex

If this code is helpful to your research, please consider citing our paper by:

@article{gao2017semi,
  title={Semi-Supervised Sparse Representation Based Classification for Face Recognition With Insufficient Labeled Samples},
  author={Gao, Yuan and Ma, Jiayi and Yuille, Alan L},
  journal={IEEE Transactions on Image Processing},
  volume={26},
  number={5},
  pages={2545--2560},
  year={2017},
  publisher={IEEE}
}