Polyhedral Object Recognition by Indexing

Radu Horaud (1, 2) , Humberto Sossa (1, 2)

1 MOVI - Modeling, localization, recognition and interpretation in computer vision
2 LIFIA - Laboratoire d'Informatique Fondamentale et d'Intelligence Artificielle

In computer vision, the indexing problem is the problem of recognizing a few objects in a large database of objects while avoiding the help of classical image-feature-to-object-feature matching paradigm. In this paper we address the problem of recognizing three-dimensional polyhedral objects from two-dimensional images by indexing. Both the objects to be recognized and the images are represented by weighted graphs. The indexing problem is therefore the problem of determining whether a graph extracted from the image is present or absent in a database of model graphs. We introduce a novel method for performing this graph indexing process based on both polynomial characterization of binary and weighted graphs, and on hashing. We describe in detail the polynomial characterization and then we show how it can be used in the context of polyhedral object recognition. Next we describe a practical recogntion by indexing system that includes the organization of the database, the representation of polyhedral objects in terms of 2D characteristic views, the representation of these views in terms of graphs, and the associated image processing. Finally, some experimental results allow the evaluation of the system performance.

Mots clés: Object recognition; Polyhedral object representation; Polynomial graph characterization; Indexing; Hashing; Feature extraction

Domaines: Synthèse d'image et réalité virtuelle [cs.GR]

Radu Horaud, Humberto Sossa. Polyhedral Object Recognition by Indexing. Pattern Recognition, 1995, 28 (12), pp.1855--1870. ⟨10.1016/0031-3203(95)00048-8⟩. ⟨inria-00590043⟩