Biological-inspired model for hybrid-order chromatic texture segregation

In this study, a computational model for chromatic texture segregation is developed. The model attempts to simulate the visual processing characteristics by mimicking the visual perception. According to Hering's opponent theory, we deal with color information in a color space with three opponent axes. The algorithm extracts 1st-order features by a Gaussian filter and 2nd-order features by a set of Gabor filters as so called Gabor wavelets. The hybrid-order features are combined at a common site to detect the boundary. The model is very intuitive and physiological relevant such that it reserves opportunities for further approaches.