In the following experiments (Grey46, Grey134, Grey209), we show how the white
noise initial converges to a blue-noise result.
In each of the 3 tests, the original grayscale input is a textureless image with
constant graylevel (46, 134, 209). The corresponding
radially averaged power spectra is shown underneath to evidence the blue-noise
property.
Grayscale input: grey 46, size 256X256
| Grey46 | Initialization (a white noise image) | after 3000 iterations | after 5000 iterations | Result |
| Test images |
|
 |
 |
 |
| The radially averaged power spectra |
 |
 |
 |
 |
Grayscale input: grey134, size 256X256
| Grey134 | Initialization (a white noise image) | after 3000 iterations | after 5000 iterations | Result |
| Test images |
|
 |
 |
 |
| The radially averaged power spectra |
 |
 |
 |
 |
Grayscale input: grey209, size 256X256
| Grey209 | Initialization (a white noise image) | after 3000 iterations | after 5000 iterations | Result |
| Test images |
|
 |
 |
 |
| The radially averaged power spectra |
 |
 |
 |
 |
