Felzenszwalb-Huttenlocher algorithm

The Felzenszwalb-Huttenlocher algorithm (English Felzenszwalb-Huttenlocher algorithm ) is an edge-oriented method for segmenting a (raster) image into objects. It was presented in 1998 by Pedro F. Felzenszwalb and Daniel P. Huttenlocher and later developed further by them.

The algorithm first places an edge between every two neighboring pixels of an image, which are weighted according to features such as the difference in brightness and color of the neighboring pixels. Image segments are then formed from each individual pixel, which are merged in such a way that the difference between the edge weightings within a segment remains as small as possible and is as large as possible between adjacent segments.

software

The algorithm is implemented in the free image processing libraries Scikit-image .

credentials

↑ Felzenszwalb, Huttenlocher: Efficiently computing a good segmentation . In: Proceedings DARPA Image Understanding Workshop . Monterey, California, USA, 1998, pp. 251-258.
↑ scikit-image.org: Module: segmentation - skimage docs. Retrieved September 8, 2018 .

Web links

Examples of image segmentation with the Felzenszwalb-Huttenlocher algorithm

[1] Felzenszwalb, Huttenlocher: Efficiently computing a good segmentation . In: Proceedings DARPA Image Understanding Workshop . Monterey, California, USA, 1998, pp. 251-258.

[2] scikit-image.org: Module: segmentation - skimage docs. Retrieved September 8, 2018 .