Open Access

Table 2

List of the twenty metrics considered.

Designation Metrics Definition
M1 Normalized density M 1 = s = 0 35 | p s ( i ) - p s ( g ) | . $ M1=\sum_{s=0}^{35}\left|{p}_s^{(i)}-{p}_s^{(g)}\right|.$Where p s ( i ) $ {p}_s^{(i)}$ is the proportion of pixels in the original image that belong to the hue sector s of the CIE L*C*h* color space. The hue sector s contains the points whose hue is between 10s and 10(s + 1). p s ( g ) $ {p}_s^{(g)}$ is the proportion for the gamut p.
M2 Maximum luminance [16] Difference between the highest luminance value in the palette and the highest luminance value in the image in absolute value.
M3 Minimum luminance Difference between the lowest luminance value in the gamut and the lowest luminance value in the image in absolute value.
M4 Delta luminance Difference between the highest and the lowest luminance values in the gamut. This value is independent of the image.
M5 Maximum chroma [16] M 5 = s = 1 6 | max ( C* ) s ( i ) - max ( C* ) s ( g ) | . $ M5=\sum_{s=1}^6\left|{\mathrm{max}\left(\mathrm{C*}\right)}_s^{(i)}-{\mathrm{max}\left({C*}\right)}_s^{(g)}\right|.$Here the sectors are chosen to be: s = 1 for h ∈ [−10, 40[(red color), s = 2 for h ∈ [40, 75[(orange color), s = 3 for h ∈ [75, 100[(yellow color), s = 4 for h ∈ [100, 200[(green color), s = 5 for h ∈ [200, 290[(blue color) and s = 6 for h ∈ [290, 350[(purple color).
M6 Minimum chroma Same as M5 but with the difference of minimum chroma in each sector for the image and the gamut.
M7 Average chroma Same as M5 but with the average chroma value rather than the maximum chroma value.
M8 Standard deviation of chroma Same as M5 but with the difference of the standard deviation values of the chroma.
M9 Convex hull area of a*b* [21] Convex hull area of the gamut when projected on the a*b* plane. This value is independent of the image.
M10 Image center of mass Euclidian distance in L*a*b* space between the barycenter of the ensembles of color points present in the gamut and image.
M11 Image colorfulness [34] Ratio of the computed Hasler colorfulness of the simulated image and the Hasler colorfulness of the original image.
M12 Image complexity [38] The complexity of the pattern within the image using the average spatial frequency of the image.
M13 Capacity to infer colors from the semantic content of the image 1 if the content of the image gives a lot of information on the expected image (e.g. color of the tree leaves, …), 0 otherwise.
M14 Pixel delta RGB volume For all pixels of the original and simulated images, the difference of the red, green and blue channels is calculated. All these differences can be plotted in a “RGB difference” solid whose convex hull has a volume that can be calculated.
M15 Pixel delta L*a*b* volume Same but the pixel differences are calculated in L*, a*, b*.
M16 Distance of the white point selected to the image gamut [39] Distance between the white point selected and the center of mass of the image gamut.
M17 Gamut centers hue difference Difference of hue values between the center of mass of the image gamut and the center of mass of palette gamut.
SSIM Structure Similarity Index Measure Comparison image structures, luminance, and contrast to assess visual similarity.
MSE Mean Square Error Measures the average squared difference between original and reproduction images as the mean squared difference (done on each channel here).
PSNR Peak Signal to Noise Ratio Measures the average squared difference between original and reproduction images.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.