Image Segmentation

This tool performs image segmentation, following the technique by Baatz and Schäpe (2000).

It has the following parameters:

  1. Compactness weight – The weight assigned to compactness. Compactness indicates the degree of clustering in an area (Zhong et al., 2005).

  2. Colour weight – The weight assigned to colour heterogeneity. The colour heterogeneity includes all channels of a multiband image (Zhong et al., 2005).

  3. Maximum allowable cost function – Only regions where the merge criterion is lower than this number will be combined into a segment. The merge criterion is given by (Zhong et al., 2005):

\[f = w_{col} h_{col}+(1-w_{col})[w_{com} h_{com}+(1-w_{com}) h_{sth}]\]

where:

  • \(w_{col}\) – weight assigned to colour heterogeneity.

  • \(h_{col}\) – colour heterogeneity.

  • \(w_{com}\) – weight assigned to circle-like compactness homogeneity.

  • \(h_{com}\) – circle-like degree of homogeneity.

  • \(h_{sth}\) - rectangle-like degree of homogeneity.

  1. Use K-Means to group segments – Segments can be group by means of a K-means classifier.

  2. Number of clusters – The number of clusters the K-means classifier must produce if that option was selected.

_images/clustseg.png

References

Baatz, M, Schäpe, A., 2000. Multiresolution segmentation: An optimization approach for high quality multi-scale image segmentation. Proceedings of Angewandte Geographische Informationsverarbeitung XII, 12-23