Multiple features data fusion method in color texture analysis

A new algorithm is developed to represent colored texture by effectively merging the texture feature, color feature, together with spatial correlation of color and texture based on incomplete tree-structured wavelet decomposition. Experiments are conducted on a set of 20 natural colored texture images in which multiple features fusion and classification performance are compared on the basis of the pyramid wavelet decomposition (PWD), Incomplete tree-structured wavelet decomposition (ICTSWD) and wavelet packet decomposition (WPD). Class correct rates of multiple features fusion based on PWD is 85.78% and class correct rates based on WPD is 91.03% with the dimensionality increased exponentially, however, the dimensionality of feature fusion based on ICTSWD is descended greatly because of selective decomposition in subband, which class correct rates is 90.63%. It is demonstrated that multiple features fusion based on ICTSWD has better classification performance and anti-noise ability than fusion based on PWD and WPD.