Invariant Characterization of the Hough Transform for Pose Estimation of Arbitrary Shapes

We develop a new formulation for including invariance in a general form of the Hough transform. We first develop a formal definition of the Hough transform mapping for arbitrary shapes and general transformations. We then include an invariant characterization of shapes and we develop and apply our technique to extract shapes under similarity and affine transformations. Our characterization does not require the computation of properties for lines or other primitives that compose a model, but is based solely on the local geometry given by points on shapes. Experimental results show that the new technique is capable of extracting arbitrary shapes under occlusion and when the image contains noise.