A graph-based approach for image segmentation that employs genetic algorithms is proposed. An image is modeled as a weighted undirected graph, where nodes correspond to pixels, and edges connect similar pixels. A fitness function, that extends the normalized cut criterion, is employed, and a new concept of nearest neighbor, that takes into account not only the spatial location of a pixel, but also the affinity with the other pixels contained in the neighborhood, is defined. Because of the locus-based representation of individuals, the method is able to partition images without the need to set the number of segments beforehand. As experimental results show, our approach is able to segment images in a number of regions that well adhere to the human visual perception. © 2012 Springer-Verlag.
An evolutionary and graph-based method for image segmentation
Amelio A.;
2012-01-01
Abstract
A graph-based approach for image segmentation that employs genetic algorithms is proposed. An image is modeled as a weighted undirected graph, where nodes correspond to pixels, and edges connect similar pixels. A fitness function, that extends the normalized cut criterion, is employed, and a new concept of nearest neighbor, that takes into account not only the spatial location of a pixel, but also the affinity with the other pixels contained in the neighborhood, is defined. Because of the locus-based representation of individuals, the method is able to partition images without the need to set the number of segments beforehand. As experimental results show, our approach is able to segment images in a number of regions that well adhere to the human visual perception. © 2012 Springer-Verlag.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
