Robust histogram generation from the HSV space based on visual colour perception

摘要

We perform a detailed analysis of the visual properties of the Hue, Saturation and intensity Value (HSV) colour space. Relative importance of hue and intensity is determined based on the results of these analyses, which helps us to suggest a weight function for estimating the degree by which any pixel is colour-dominant or greyscale-dominant. We construct a histogram in which each pixel contributes to a grey component as well as to a hue component. The histogram is robust even in the presence of a considerable extent of noise in the input image.