Abstract 

The endmembers of a hyperspectral image (HSI) are more likely to be generated by independent sources and be mixed in a macroscopic degree before arriving at the sensor element of the imaging spectrometer as mixed spectra. The paper titled "Constrained Nonnegative Matrix Factorization for Blind Hyperspectral Unmixing incorporating Endmember Independence" presents a novel blind HU algorithm, referred to as Kurtosis-based Smooth Nonnegative Matrix Factorization (KbSNMF) which incorporates a novel constraint based on the statistical independence of the probability density functions of endmember spectra. The proposed algorithm manages to outperform several state of the art NMF-based algorithms in terms of extracting endmember spectra from hyperspectral data. The attached datasets are utilized to reproduce the results presented in the above-mentioned paper.

Instructions: 

The experiments in the paper titled "Constrained Nonnegative Matrix Factorization for Blind Hyperspectral Unmixing incorporating Endmember Independence" utilizes a set of different hyperspectral datasets with varying parameters such as the no. of endmembers, no. of spectral bands, no. of pixels, and SNR level. The attached datasets are in .mat format and utilize the following nomenclature.

 

"endmembers_simulated_dataset_XX.mat", where XX denotes the number of endmembers

"spectral_simulated_dataset_YY.mat", where YY denotes the number of spectral bands

"pixels_simulated_dataset_ZZ.mat", where ZZ denotes the number of pixels

"noisy_simulated_dataset_SNR_SS.mat", where SS denotes the SNR level

"dataset_simulated.mat" is utilized for general performance evaluation and parameter selection

"dataset_real_samson.mat" is utilized for general performance evaluation of real data