DISTANCE AND MEDIAL AXIS TRANSFORMS

The Distance Transform transforms a binary image in an integer (usually byte is sufficient) image. Each pixel of the "foreground" is labelled with its distance from the boundary.

The algorithm is as follows:

The Medial Axis Transform aims to find the ridges in the distance-transform image. The idea is to expand a circle around each pixel (with positive value) until it touches the boundary. If the circle touches the boundary in "more than one point", the pixel is on the medial axis. An algorithm is:

The outcome of the medial axis transform can be refined by thinning the ridges, for example removing noise spur of short length.


G. Borgefors, Distance transformations in arbitrary dimensions, Computer Vision Graph. Image Process. 27, 1984, 321-345


Marco Corvi - Page hosted by geocities.com.