We now apply our scheme given by Eqn. 9 to the problem of
binary images with noise. Since we are looking at black and
white images, where 0 corresponds to black and 255 to white,
the threshold value 
is taken as 
rather than
0. In Figure 8, we add noise to a
black and white image of a hand-written character. The noise is
added as follows; 
noise means that at 
of the pixels,
we replace the given value with a number chosen with uniform distribution
between 0 and 255.
Thus, a full spectrum of gray noise is added to the original binary image,
The left column give the original figure with the corresponding percentage
of noise; the right column are reconstructed values. We stress once
again that the figures on the right are converged; they stop automatically,
and continued application of the scheme yields no change in the results.
Results are reconstructed from 
, 
, 
, and 
noise.
Figure 8: Image restoration of Binary Images with Grey-Scale Salt-and-Pepper
Noise Using Min/Max Flow