Abstract 
I will talkt about the Polar Distance Transform (PDT), which is a weighted distance transform where local steps in the radial direction, with respect to a given origin, have a different (higher) weight than tangential steps. In this way, circular paths will be preferred over paths in the radial direction. The PDT can be combined with a cost function to create the Grey Weighted Polar Distance Transform (GWPDT). Here, the length of a local step is multiplied with the value of a cost function image, so that paths along low values are shorter than paths along high values
The GWPDT is a useful tool for finding shortest paths corresponding to approximately circular shapes in images.
