Title :
Color texture feature selection by MIFS for image classification
Author :
Maia, H.E. ; Hammouch, A. ; Bakrim, M´Hamed
Author_Institution :
LRIT Unite associe au, Univ. Mohammed V-Agdal, Rabat, Morocco
fDate :
Sept. 30 2010-Oct. 2 2010
Abstract :
The work presented in this article consists in evaluating the contribution of the use of a texture-color combination in forest mapping. We transform a Spot image into different color spaces, then we extract nine parameters using Laws filter. The resulting set of parameters is reduced using MIFS algorithm based on mutual information. The results are promising and give satisfying classification results.
Keywords :
feature extraction; filtering theory; image classification; image colour analysis; image texture; set theory; Laws filter; MIFS; color texture; feature selection; forest mapping; image classification; spot image; Accuracy; Feature extraction; Image color analysis; Image segmentation; Mutual information; Support vector machines; Training; Classification; Color; Laws filter; MIFS; Texture;
Conference_Titel :
I/V Communications and Mobile Network (ISVC), 2010 5th International Symposium on
Conference_Location :
Rabat
Print_ISBN :
978-1-4244-5996-4
DOI :
10.1109/ISVC.2010.5656251
