Shape modeling of multiple objects from shading images using genetic algorithms

This paper describes about an application of genetic algorithms (GAs) to modeling of multiple object from CCD images. Shape modeling is a very important issue for shape recognition for robot vision, representing 3D shapes in the virtual world, and so on. Superquadrics are often used for shape modeling because they can represent various shapes by using a single equation. In this paper, we propose a new method for applying GAs to estimation of the superquadrics parameters of every objects in a shading image which are taken with a CCD camera. The superquadrics parameters are represented by strings. The string is evaluated by the similarity between the given 2D shading image and the calculated shading image from the 3D shape represented by the parameters. For finding the model parameters of each object in the image, sharing scheme is employed so that multiple solutions can be held in the population of the strings. Some results of the computer experiments demonstrate that the proposed method can provide good model descriptions of the 3D object in shading images.