feng135799 / componentspd Goto Github PK

Component symmetric positive definite matrices (CSPD) matrices for describing image sets.
CSPD: A low-dimensional discriminative data descriptor(a special type of improved covariance descriptors) than traditional covariance descriptors for image set classification.

Written by Kai-Xuan Chen (e-mail: [email protected])
version 2.0 -- December/2018
version 1.0 -- June/2017

Please cite the following paper (more theoretical and technical details) if your are using this code:

Kai-Xuan Chen, Xiao-Jun Wu. Component SPD matrices: A low-dimensional discriminative data descriptor for image set classification[J]. Computational Visual Media, 2018, 4(3): 245-252.

BibTex :

@article{Chen2018Component,  
      title={Component SPD matrices: A low-dimensional discriminative data descriptor for image set classification},  
      author={Chen, Kai-Xuan and Wu, Xiao-Jun},  
      journal={Computational Visual Media},  
      volume={4},  
      number={3},  
      pages={245--252},  
      year={2018},  
      publisher={Springer}  
}  

The ETH-80 dataset is needed to be downloaded(https://github.com/Kai-Xuan/ETH-80/),
and put 8 filefolders(visual image sets from 8 different categories) into filefolder '.\ETH-80'.
Please run 'read_ETH.m' to generate CSPD matrices. Then run 'run_ETH.m' for image set classification.

For classification, we employ four NN classifiers and four discriminative classifiers in this source code(Version 2.0).

Ker-SVM : Qilong Wang implemented a one-vs-all classifier by using LIBSVM package in the paper:
Q. Wang, P. Li, W. Zuo, and L. Zhang. Raid-g: Robust estimation of approximate infinite dimensional gaussian with application to material recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4433-4441, 2016.
Chih-Chung Chang and Chih-Jen Lin. Libsvm: a library for support vector machines. ACM transactions on intelligent systems and technology (TIST), 2(3):27, 2011.

LogEKSR : This method was proposed in the paper:
P. Li, Q. Wang, W. Zuo, and L. Zhang. Log-euclidean kernels for sparse representation and dictionary learning. In Proceedings of the IEEE International Conference on Computer Vision, pages 1601-1608, 2013.

COV-LDA/COV-PLS : This method was proposed in the paper:
R. Wang, H. Guo, L. S. Davis, and Q. Dai. Covariance discriminative learning: A natural and efficient approach to image set classification. In Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on, pages 2496-2503. IEEE, 2012.