As shown in the example of Figure 6-11, the implementation of the sigma and extreme value filters improves the quality of the Laplacian zero crossings and thus also the quality of edges. It is important to stress that these two filters improve the quality of the image so strongly that it becomes possible to apply the simplest method of edge detection. This is the binarization of the gradient of the gray values:
Gradx(x, y) = GV(x, y) – GV(x – 1, y)
Grady(x, y) = GV x, y) – GV(x, y – 1).
Here GV(x, y) is the gray value of the pixel (x, y). The values ...