Dual Graph Contraction with LEDA

摘要

Graphs are useful tools for modeling problems that occur in a variety of fields. In machine vision graph based solutions have been successfully applied to many image processing problems e.g. quad trees for image compression and region adjacency graphs for segmentation. The application of graphs to machine vision problems poses special problems due to the underlying size of the image e.g. a graph representing the base level of a 512x512 image has over 200.000 nodes. The large size of the graphs make issues of both space and time complexity important when designing algorithms for ma- chine vision problems.