In addition to clustering, a fuzzy approach to image segmentation is proposed in this work also for the handling of spatial information of topological type. Topological information is taken into account, extending the fuzzy connectivity, to define a segmentation method of grey level images. The main innovative aspects of the proposed method consist in the automatic processing of real images, without using any parameter and fixed threshold, and in the simplicity and speed of the computation, realised through simple fuzzy operators and a noniterative growing mechanism. Quantitative evaluation of preliminary results, obtained on real images from various application domains, proves the feasibility of the method
A segmentation method based on fuzzy topology and clusteringProceedings of 13th International Conference on Pattern Recognition
DELLEPIANE, SILVANA
1996-01-01
Abstract
In addition to clustering, a fuzzy approach to image segmentation is proposed in this work also for the handling of spatial information of topological type. Topological information is taken into account, extending the fuzzy connectivity, to define a segmentation method of grey level images. The main innovative aspects of the proposed method consist in the automatic processing of real images, without using any parameter and fixed threshold, and in the simplicity and speed of the computation, realised through simple fuzzy operators and a noniterative growing mechanism. Quantitative evaluation of preliminary results, obtained on real images from various application domains, proves the feasibility of the methodI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.