Hue and Saturation Dispersion Modeling. Application to the Improvement of the Segmentation in the HS Sub-space

In this paper, a modeling of the hue and saturation dispersions of classes projected in the HS plane is presented. The relevance of the proposed modeling is its application in general clustering processes in the HS domain. In a previous work about class separation in the HS plane for segmentation improvement, it was necessary to calculate the hue and saturation variances of the classes for their different locations in the HS plane. Classes may be composed by a high number of samples and variances are distinct for the different locations within the HS plane, so this process needs a high processing calculus time. In the current paper, the hue and saturation dispersions are modeled (requiring a lower processing time) and are applied to that segmentation process. The authors propose to use the geometric-analytical relationships between the HSI space and the YC₁C₂ tristimulus. These relationships allow to estimate the dispersions of the classes projected in the HS plane from their dispersions in the C₁C₂ linear chromatic plane. The results obtained in different practical applications show the validity of the proposal and a remarkable improvement in the processing time.