Bond Dissociation Energy and Pattern Spectrum for Shape Classification

Abstract

The paper represents an effective, robust and shape-preserving, pattern spectrum based on sigma-pi model. A new idea behind the translation, rotation and scale invariant features are introduced called bond-dissociation energy of sigma and pi bond in a sigma-pi model. In this model, we map the structure of pixels in a shape to the structure of atoms in a crystalline solid-state substance. The model compute the energy required to break the atom from a lattice. The sigma-pi model corresponding to the bond-dissociation energy pattern spectrum provides excellent power, which is demonstrated by excellent retrieval performance on several popular shapes benchmarks, including MPEG-7 CE-Shape-l data set, Myth dataset, Tool dataset and Kimia’s data set. Experimental results obtained from popular databases demonstrate that the proposed model can achieve comparably better results than existing algorithms.