Awesome

Hyperspectral Unmixing using Transformer Network

Preetam Ghosh, Swalpa Kumar Roy, Bikram Koirala, Behnood Rasti, and Paul Scheunders

<strong>:fire:New:bangbang:</strong></font></sup> Code is now available here.

The repository contains the PyTorch implementations for Hyperspectral Unmixing using Transformer Network.

Dataset

• Simulated Dataset of 80$\times$80 pixels (see Fig. \ref{Image and Endmembers} (a)) is generated by the linear combination of three endmembers (i.e., Iron (Fe$_2$O$_3$), Silica (SiO$_2$), and Calcium (CaO)) (see Fig. \ref{Image and Endmembers}(b)). Each hyperspectral pixel contains reflection values for 200 different bands covering the wavelength range [1000-2500] nm. This image contains 16 squares of 20 $\times$ 20 pixels with different ternary mixtures (see the first column of Fig. \ref{fig:Sim_Abun})}

If you use the code in your research, we would appreciate a citation to the original paper:

@article{ghosh2019hyperspectral,
    	title={Hyperspectral Unmixing using Transformer Network},
	author={Ghosh, Preetam and Roy, Swalpa Kumar and Koirala, Bikram and Rasti, Behnood and Scheunders, Paul},
	journal={IEEE Transaction on Geoscience and Remote Sensing},
	volume={60},
	no.={1},
	pp.={01-16},
	year={2022}
	}