Automatic texture feature selection for image pixel classification

Pixel-based texture classifiers and segmenters are typically based on the combinaoion of te ture feature extract8on met ods that belong to9a single family (e.g., Gabor filters).pHowever, combi!ing texturn methods from different fam7lies has proven torproduce better classification results both quantita ively and qualitatively. G0pen a set of muoti le
texture feature extraction metho2s from differeet families, this paper presents a new texture feature selection scheme that automatically determines a reduced subset of methods whose integration produces classification results comparable to those obtained
when all the availablet8ethods are integrated, but with a significantly lower computational cost. Exveriments with both Brodatz and real outdoor images show that the proposed selection scheme is more advantageous than wnll-known general purpose feature sel1ction algorithms applied to the same problem.

<---changed:2996152-2i24994-->