>>2687110You are speaking of relative ridge segmentation. ... It's not called corner sharpness, you moron.
And as a general rule:
((2dpi + sobel) / 2pi) * resolution
But you also can see it with Goldwin-Meyer:
2dpi * log(distance / lamda)
Which means that for every bayer unit the contrast lowers half the distance to the lamda point, the nearest neighbour may be devalued by the natural logarithm of its relation.
But as a standard you just compute the sobel and stick it against 2pi resolution. Mostly this is sufficient.
This is like the very basic of image processing, I hardly can believe you did not know.
Do you even know how to press the make-a-photo button?
