a la physique de l'information - Lisa - Université d'Angers
a la physique de l'information - Lisa - Université d'Angers
a la physique de l'information - Lisa - Université d'Angers
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
8<br />
Fig. 7 Color histogram in the RGB colorimetric cube [0, 255] 3 for image “Zelda” (left) and<br />
image “Fruits” (right) from Fig. 6.<br />
variations of some generic reference colors, <strong>la</strong>rge scales are re<strong>la</strong>ted to the several <strong>la</strong>rgely<br />
distinct colors that frequently compose natural images. And these different colors are<br />
usually recruited in a self-simi<strong>la</strong>r way across scales, as manifested by Fig. 8.<br />
The fractal property of the colorimetric organization of natural images, as manifested<br />
by Fig. 8, complements other fractal properties of the colors recently reported in<br />
[3,2]. The results of Fig. 8 with the box-counting method, characterize, as we already<br />
mentioned, the support of the three-dimensional histogram. Equivalently, this measures<br />
which colors are used and which are not used in the image. This support, according<br />
to Fig. 8, tends to disp<strong>la</strong>y a fractal structure, with clusters and voids spanning many<br />
scales. This is i<strong>de</strong>ntified, from the slopes in Fig. 8, by a fractal dimension D which is<br />
known as the box-counting or capacity dimension, or dimension of the support of the<br />
distribution [19,24]. By contrast, references [3,2] measure the corre<strong>la</strong>tion dimension of<br />
the three-dimensional histograms, with two different estimators in [3] and in [2]. This is<br />
a distinct dimension compared to the capacity dimension measured here. The capacity<br />
dimension and the corre<strong>la</strong>tion dimension are two important instances in the infinite series<br />
of fractal dimensions which can be <strong>de</strong>fined for fractal structures [13,24]. In general,<br />
the capacity dimension is ≥ to the corre<strong>la</strong>tion dimension, and this is in<strong>de</strong>ed verified<br />
for color histograms here and in [3,2]. The fractal characterization of [3,2], via the corre<strong>la</strong>tion<br />
dimension, characterizes simultaneously, in a joint manner, both the support<br />
of the histogram and the popu<strong>la</strong>tions in the histogram. Equivalently, this measures simultaneously<br />
which colors are used in the histograms, and which popu<strong>la</strong>tions of pixels<br />
distribute among these occupied colors. A fractal organization is found in [3,2], for the<br />
way the pixels popu<strong>la</strong>te the occupied colorimetric cells of the histogram. The results<br />
of Fig. 8 here, reveal that the occupied colorimetric cells alone, irrespective of their<br />
popu<strong>la</strong>tions, disp<strong>la</strong>y a fractal organization. These are two distinct fractal properties,<br />
which are both useful to characterize the complex colorimetric organization of natural<br />
images.<br />
A consistent behavior of the capacity dimension can be observed when the boxcounting<br />
procedure is applied to natural gray images co<strong>de</strong>d as RGB colors images with<br />
three i<strong>de</strong>ntical color components R = G = B. In this case, as shown in Fig. 9, the num-<br />
174/197