dunazo / endnet

EndNet: Sparse AutoEncoder Network for Endmember Extraction and Hyperspectral Unmixing

• Download datasets from https://drive.google.com/open?id=0B1kqE05deSqxYm8wS1F3UXZYeWM
• Extract datasets.zip to project directory.
• Run demo.m to visualize results for a selected dataset.

Code will be released soon on this page. Stay tuned!...

Our Results

Samson Data and Abundance Map Results

Jasper Data and Abundance Map Results

Urban Data and Abundance Map Results

Cuprite Data and Abundance Map Results

Pavia University Data and Abundance Map Results

Pavia Center Data and Abundance Map Results

Gulfport Data and Abundance Map Results

IEEE 2013 Challenge Data and Abundance Map Results

DC Data and Abundance Map Results

References

Please cite the following paper:

[1] Savas Ozkan, Berk Kaya and Gozde Bozdagi Akar, EndNet: Sparse AutoEncoder Network for Endmember Extraction and Hyperspectral Unmixing, arXiv:1708.01894:

@article{ozkan17endnet,
  title={EndNet: Sparse AutoEncoder Network for Endmember Extraction and Hyperspectral Unmixing},
  author={Ozkan, Savas and Kaya, Berk and Akar, Gozde B},
  journal={arXiv:1708.01894},
  year={2017},
}