Vilaplana V, Giró-i-Nieto X, Salembier P, Marqués F. Region-based extraction and analysis of visual objects information. In Fourth International Workshop on Content-Based Multimedia Indexing, CBMI 2005. Riga, Letonia; 2005.  (225.03 KB)

Abstract

In this paper, we propose a strategy to detect objects from still images that relies on combining two types of models: a perceptual and a structural model. The algorithms that are proposed for both types of models make use of a regionbased description of the image relying on a Binary Partition Tree. Perceptual models link the low-level signal description with semantic classes of limited variability. Structural models represent the common structure of all instances by decomposing the semantic object into simpler objects and by defining the relations between them using a DescriptionGraph.