Bug in code that computes the weighs based on the mask

[2006pmach]

Hi,

I noticed a bug in you implementation of the function def weight_context(self, mask):
Although you claim in the comment that the

#sliding window size is 7

you use in fact a window size of 6 as indexing an array in numpy with array[i-3:i+3] returns the values at i-3, i-2,i-1, i ,i+1, i+2 and does not include i+3. Thus, you are using a window size of 6 which is not centrally symmetric and puts different weights for pixels/holes on the right than on the left or bellow than above. So just add the +1 to all the corresponding lines...