Title :
Electrostatic formulation for adaptative dilation
Author :
Lavialle, Olivier ; Baylou, Pierre
Author_Institution :
Equipe Signal/Image ENSERB, Talence, France
Abstract :
We introduce a new concept of caricature in order to exacerbate the main morphological characteristics of objects. Here, the term caricature implies that we are looking for a method leading to a simplification or an exaggeration of the particularities of objects. Our technique is based on the shape of the object considered as an insulator charged with unipolar electricity. The electric forces exerted on each pixel belonging to the contour create the desired transformation. The contextual information is also taken into account through the definition of attractive forces between objects. We consider only the case of the dilatation of objects in a binary image. The theory is illustrated through a problem concerning connection of objects in a binary image and through an example of contour closing.
Keywords :
electrostatics; image processing; adaptative dilation; binary image; contour closing; electric force; electrostatic formulation; insulator; unipolar electricity; Extremities; Force; Image analysis; Image edge detection; Insulators; Permittivity; Shape;
Conference_Titel :
Signal Processing Conference (EUSIPCO 1998), 9th European
Conference_Location :
Rhodes
Print_ISBN :
978-960-7620-06-4