Matching and Clustering: two Steps towards Automatic Object Model Generation in Computer Vision

Abstract

In this paper, we present a general framework for a system for automatic model generation and recognition of 3D polyhedral objects. Such a system has many applications ill robotics: recognition, localization, grasping... Here we focus upon one question: from multiple images of one 3D object taken from different unknown viewpoints, how to recognize those which represent identical aspects of tile object? Briefly, is it possible Io (leternline automatically if two images are similar or not? The two stages detailed ill this paper are matching of two images and clustering of a set of images. Matching consists of finding the common features of two images while no information is known about the image content, camera motion or calibration. Clustering consists of regrouping into sets tile images representing identical aspects of the objects. For both stages, experimental results on real images are shown.