Nature scene image extraction method based on HVS

A nature scene image color-texture feature extraction method based on human visual system(HVS) was proposed.Two types of low-level image features color and texture feature extraction methods were introduced,and adaptive dominant colors principle described color features,spatial and frequency characterristic of texture grayscale component describes texture features,using adaptive clustering algorithm(ACA) to get adaptive dominant colors and steerable pyramid filter decomposition to extract texture features.Meanwhile,optimal colorcomposition distance(OCCD) and texture distance metric approach were also presented.Finally experiment results and analysis were given out.