An unified interface for benchmarking HSI denoising algorithms on various datasets under different noise settings.
-
Download the library of HSI denoising algorithms from OneDrive and put the
lib
directory in thematlab
folder. -
Type
addpath(genpath('lib'));
in matlab command line; now you can use any algorithms with an unified interface defined indemo_fun.m
. For instance, you can useBM4D
to denoise a hyperspectral image viademo_fun(noisy_hsi, sigma_ratio, 'BM4D')
-
Benchmark your selected algorithm on the whole dataset by
Main_Gauss/Complex/Real.m
; Get the quantitative results viaResult_Gauss/Complex.m
The bibtex of the collected algorithms can be found in benchmarks.bib
- The code of different HSI denoising algorithms is collected online.
- Special thanks to these authors for making their source code publicly available.
Adding your own algorithms is very easy: Simply put your source code in lib
directory, then write your algorithm interface in demo_fun.m
.