Considering that humans are the judge of what constitutes as a well spaced font, one should look at the workings of the human brain. It turnes out we already kn

Considering that humans are the judge of what constitutes as a well spaced font, one should look at the workings of the human brain. It turnes out we already know the algorithm employed by the visual cortex for edge detection, the _Gabor filter_ http://authors.library.caltech.edu/2653/1/MALjosaa90.pdf http://authors.library.caltech.edu/2653/1/MALjosaa90.pdf Also Wikipedia has a fantastic overview for the Gabor filter: http://en.wikipedia.org/wiki/Gabor_filter http://en.wikipedia.org/wiki/Gabor_filter Of the approaches mentioned in the original article, "wavelet" masks, are the ones closest to the calculation done in the V1. A quick search on http://scholar.google.com http://scholar.google.com for "Gabor space visual cortex" uncovers that much research was dedicated to this topic.