SSN

Code for Spatio-Spectral Networks (SSN), a method to compute color-texture descriptors from images, from the paper "Spatio-spectral networks for color-texture analysis".

Preprint: https://arxiv.org/abs/1909.06446

Journal version: https://www.sciencedirect.com/science/article/pii/S0020025519310874

The code was implemented using Matlab and C++. The binary "mexa64" works on Linux 64 bit systems, while "mexw64" works on Windows 64 bits. If you are running a different system, you may need to compile the C++ sources ".cpp" using MEX:

mex('SSN_getFeatureMaps.cpp')
mex('SSNgrey_getFeatureMaps.cpp')

Usage

• Given a pixel matrix representing an 'image' and a maximum radius 'rn' for neighboring pixel analysis, call the main function SSN:

features = SSN(image, rn)

• It works either with 3d matrices (i.e., color images) or 2d matrices (grayscale). However, notice that the method was developed for color images (see the paper).

Cite

If you use this method, please cite our paper:

Scabini, Leonardo FS, Lucas C. Ribas, and Odemir M. Bruno. "Spatio-spectral networks for color-texture analysis." Information Sciences 515 (2020): 64-79.

@article{scabini2020spatio,
  title={Spatio-spectral networks for color-texture analysis},
  author={Scabini, Leonardo FS and Ribas, Lucas C and Bruno, Odemir M},
  journal={Information Sciences},
  volume={515},
  pages={64--79},
  year={2020},
  publisher={Elsevier}
}