GLCM_Features(glcm)

Version 1.4.0.0 (3,13 ko) par Pawel Kleczek
Vectorized GLCM texture features calculations
0.0
(0)
1,4K téléchargements
Mise à jour 25 jan. 2017

Afficher la licence

This code is a vectorized version of the code submitted by Avinash Uppuluri.
The speedup (tested for the same subset of features) for a 200x200x4 GLCM matrix is about:
- 300x with respect to the original, non-vectorized version
- 10x with respect to the Avinash Uppuluri's own vectorized version
- 7x with respect to the vectorized implementation by Patrik Brynolfsson
NOTE: Formulas of features used in this implementation are based solely on the paper by Haralick et al.:
Haralick RM, Shanmuga K, Dinstein I.: Textural features for image classification. IEEE Trans Syst Man Cybern 3: 610-621, 1973
Other implementations may use slightly different formulations of "Sum of square: variance", "sum variance" and "difference variance". Always make sure which version is more suitable in the given case.

Citation pour cette source

Pawel Kleczek (2025). GLCM_Features(glcm) (https://fr.mathworks.com/matlabcentral/fileexchange/56661-glcm_features-glcm), MATLAB Central File Exchange. Extrait(e) le .

Compatibilité avec les versions de MATLAB
Créé avec R2014a
Compatible avec toutes les versions
Plateformes compatibles
Windows macOS Linux
Tags Ajouter des tags
Remerciements

Inspiré par : GLCM texture features

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Version Publié le Notes de version
1.4.0.0

Only changed the description
1.3.0.0

- Fixed a bug related to hard-coded 3rd dimension of a GLCM matrix (was 4, now is size(GLCM,3)).
- Added a demo file to show how to use the function.
1.2.0.0

Fixed bug in computation of `t` for `svarh`: length(p_xplusy) gave incorrect results if `p_xplusy` was a 3x1x4 matrix (i.e. when GLCM matrix was a 2x2xk matrix).
1.1.0.0

I've ensured that the current implementation of "Sum of square: variance", "sum variance" and "difference variance" conforms to the original paper by Haralick et al.
1.0.0.0