The spectral total variation denoising algorithm

A denoising algorithm for hyperspectral images

This package provides the MATLAB codes for the spectral total variation (STV) denoising algorithm [1], which is a new denoising algorithm for hyperspectral images that estimates different noise levels across the spectral axis from observed data.
The command for the STV denoising algorithm is

out_stv = spectral_tv(hyper_noisy, opts);

where hyper_noisy is the input image and opts is the parameter. The input image should be a 3-D noisy image (hyperspectral image or video). Also, prior to writing the command, opts.beta needs to be set to [1 1 0.1]. The output image is stored as out_stv.f. Please see the attached user guide for details.

For additional information and citations, please refer to:

[1] Chien-Sheng Liao, Joon Hee Choi, Delong Zhang, Stanley H. Chan and Ji-Xin Cheng, "Denoising Stimulated Raman Spectroscopic Images by Total Variation Minimization," Journal of Physical Chemistry C, Jul. 2015.