This software applies the DSC subspace clustering algorithm [arXiv:1706.03860] to the face clustering
problem. The code uses Extended Yale B dataset which contains 64 images for each of 38 individuals in
frontal view and different illumination conditions.
Remarks:
1- If you use this code in your research/work, please cite the following papers:
@article{rahmani2017direction,
title={A Direction Search and Spectral Clustering Based Approach to Subspace Clustering},
author={Rahmani, Mostafa and Atia, George},
journal={arXiv preprint arXiv:1706.03860},
year={2017}
}
@article{rahmani2015innovation,
title={Innovation pursuit: A new approach to subspace clustering},
author={Rahmani, Mostafa and Atia, George},
journal={ICML 2017, arXiv preprint arXiv:1512.00907},
year={2015}
}
2- The provided code is not the efficient implementation of algorithm. The code provides an expressive
implementation of the algorithm and it is provided for educational purposes. If the user intends to
measure the complexity of DSC or its running time, an efficient implementation should be used.
3- The selected parameters are not the optimal parameter for any possible application. However, I
found them pretty effective in all my experiments.
4 – The DSC iterative solver uses random initialization for some of the variables.