Colour segmentation based on separate anisotropic diffusion of chromatic and achromatic channels

The paper presents a new technique for segmenting images only on the basis of colour information. It is shown how segmentation can benefit from splitting colour signals into chromatic and achromatic channels and separately smoothing them through anisotropic diffusion. Operatively, this is accomplished through two independent diffusion processes: one involves only the chromatic information, conveniently embedded in a complex function, while the other affects the lightness information. The results of the two diffusions are separately segmented by k-means clustering techniques and their combination yields the final image partition into homogeneous regions. Some experimental results are reported which verify the effectiveness of the proposed technique.