Symbol Recognition based on a Multiresolution Analysis of the Radon Transform
Abstract
This paper presents a novel approach to define a multiresolution descriptor of shapes based on the Radon transform. The key idea is to project a binary shape in the Radon space for different levels of resolution provided by a wavelet decomposition. It gives rise to a representation having a nice behaviour with respect to common geometrical transformations. The accuracy and the efficiency of the proposed algorithm in the presence of a variety of transformations is demonstrated within a shape recognition process.