Target classification using curvature scale spaces
HEDLEY C. MORRIS, Monica M. De Pass · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2006
In this paper, we present a study of target recognition based on curvature measures. The goal of this work is to examine identification algorithms based on contour and surface curvature estimations. F. Costa and R. M. Cesar have developed fractal and curvature analysis methods for the contours in 2D images. In particular, they have introduced the notion of the curvegram of a contour. This provides a visual representation of how the curvature changes with scale. We propose an extension to target recognition based on curvature descriptors but involving the whole target set not just the boundary.