A probabilistic contour is a specialization of a radial contour. When associated with a probabilistic shape model, the contour stores both a mean radial position and an associated variance for each vertex.
In this implementation, explicit upper and lower bounds for each radial are also computed in a probabilistic model. These values are computed to lie at some user-definable number of standard devaiations away from the expected location of each vertex. The bounds are used to give visual feedback about the variability in each vertex's position and also to limit the region in which edge detection will take place during image segmentation.