Propagating Segmentations
Fernand Meyer · 2019
This chapter shows how an image with sufficient contrasted objects may be segmented region-by-region and how to extract the neighboring catchment zone, separated from the previous one by the lowest pass point. It presents an example of an image with well-marked transitions between structures and substructures in the image, and without noise. The intensity of the transition between a newly segmented catchment zone and the portion of the image segmented previously is measured by the altitude of the pass point between the two regions. The transitions between neighboring catchment zones do not have the same intensity. Some zones are extremely similar in the initial image, and the gradient between them has a low value. The chapter also shows how it is possible to adapt the step-by-step segmentation method to marker-based segmentation.