MST Segmentation

The minimum spanning tree (MST) pyramid is a hierarchical image segmentation method. Its input is a weighted neighborhood graph (NG), associating a vertex to each image pixel and connecting two neighboring vertices by an edge weighted with the distance of the two pixel values in some feature space (e.g. gray scale or RGB color). The MST pyramid method pro-duces a hierarchy of partitions by using the minimum spanning tree (MST) algorithm by Boruvka and region internal/external contrast concepts. This segmentation method is well-suited for parallel processing, as the decisions in the merging process of the regions are taken locally. The results of our segmentation method are comparable to the well-known Normalized Cut segmentation. Related publications [21,14,24,59]. (see publications).

Last modified: May 07, 2010 by Nicole M. Artner