Convergency tests with white-noise initialization
The following experiments show how the optimization converges even when it starts naively from white noise.
Grayscale input |
|
|
|
Gradual sequences |
|
|
|