Color texture retrieval using wavelet decomposition in the independent components color space

In this paper, we propose a color texture retrieval method using wavelet decomposition in the independent component color space. In color texture retrieval, the product of low dimensional marginal distributions of wavelet coefficients from different color layers is preferred to substitute or approximates the high dimensional joint distributions in order to avoid the curse of dimensionality. However, the RGB color spaces is a highly correlated color space and the extracted wavelet coefficients from different layers are also correlated, which means such a substitution or approximation will not be adequate. To solve the problem, we use independent component analysis to decorrelate the R, G and B layers into three new independent layers before applying wavelet decomposition on the color texture images. Experimental results show the proposed color texture retrieval method has a retrieval rate of 82.73%, while its RGB based counterpart that ignores the inter-layer correlation has a retrieval rate of 71.26%. Theoretically, our method will also have lower computataional demands than other color texture retrieval methods, which employ additional inter-layer correlation feature descriptors or hidden Markov model(HMM) in their algorithms.