作者: Jesús Chamorro-Martínez , Belén Prados-Suárez , Daniel Sánchez
DOI:
关键词: Segmentation-based object categorization 、 Computer science 、 Minimum spanning tree-based segmentation 、 Artificial intelligence 、 Homogeneity (statistics) 、 Region growing 、 Scale-space segmentation 、 Image segmentation 、 Fuzzy logic 、 Image texture 、 Pattern recognition
摘要: This paper is the continuation of a previous work were we proposed use t-norms to measure path homogeneity in path-based fuzzy image segmentation. Here, introduce an approach automatically select function, taking into account characteristics region be segmented and its contour. To this purpose firstly approximate characteristic surrounding seed, on basis study set fixed paths. Secondly, relationship between parameter that controls behavior functions based Weber’s t-norm. Finally, obtain function from value computed slope, gives whose associated suitable model region.