# TF_LDA

This is an official implementation of the paper "C2DNDA : A Deep Framework for Nonlinear Dimensionality Reduction".

## Some Essential Information:

- Tensrflow 1.4
- CUDA 8.0
- Python 2.7


## Simple Instruction:

"data_process.py" is used for data preprocessing.

Run "tf-lda.py" to train and test.

## datasets:

Please download datasets from the following websites:

1. http://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz
2. http://yann.lecun.com/exdb/mnist/

## Copyright and Citation:

This code is ONLY released for academic use. Please do not further distribute the code (including the download link) on the public website. 
Please kindly cite our papers if you use our code in your research. Thanks and hope you will benefit from our code. 

@article{Wang2020C2DNDA, 
  author={Wang, Qi and Qin, Zequn and Nie, Feiping and Li, Xuelong}, 
  journal={IEEE Transactions on Industrial Electronics}, 
  title={C2DNDA : A Deep Framework for Nonlinear Dimensionality Reduction}, 
  DOI = {10.1109/TIE.2020.2969072},
  year={2020},
}

@inproceedings{Wang2017Convolutional,
  title={Convolutional 2D LDA for Nonlinear Dimensionality Reduction},
  author={Wang, Qi and Qin, Zequn and Nie, Feiping and Yuan, Yuan},
  booktitle={Twenty-Sixth International Joint Conference on Artificial Intelligence},
  year={2017},
  pages={2929-2935}, 
}