Computing the object colour solid using spherical sampling

+ + ⋯ + = 0 where , , ..., are the set of arbitrary real numbers, where at least one of them is not equal to zero and λ are the colour system spectra which are the products of sensor spectral sensitivities λ and an illuminant spectrum e λ i.e. s λ = λ λ . Here, we observe that the components of vector k have the geometrical meaning i.e. they constitute the normal vector parametrising the surface of the object colour solid. Because the OCS is convex, in the direction k, we can, in closed form, find the unique system response which is maximum. And, from convexity it follows we can find all points on the OCS by extremizing all directions. Formally, we propose the parametric representation with respect to k of the surface of the object colour solid. For a reflectance function λ the colour system responses are: = λ λ d , = 1,2, ... , #

[1]  Alexander D. Logvinenko,et al.  Metamer Mismatching , 2014, IEEE Transactions on Image Processing.

[2]  A. Logvinenko An object-color space. , 2009, Journal of vision.

[3]  Brian V. Funt,et al.  XYZ to ADL: Calculating Logvinenko's Object Color Coordinates , 2010, Color Imaging Conference.

[4]  Graham D. Finlayson,et al.  Making the calculation of Logvinenko's coordinates easy , 2012, Color Imaging Conference.

[5]  Sabine Süsstrunk,et al.  Spectral sharpening by spherical sampling. , 2012, Journal of the Optical Society of America. A, Optics, image science, and vision.