We investigate the problem of facial expression recognition using 3D data.Building from one of the most successful frameworks for facial analysis usingexclusively 3D geometry, we extend the analysis from a curve-basedrepresentation into a spectral representation, which allows a completedescription of the underlying surface that can be further tuned to the desiredlevel of detail. Spectral representations are based on the decomposition of thegeometry in its spatial frequency components, much like a Fourier transform,which are related to intrinsic characteristics of the surface. In this work, wepropose the use of Graph Laplacian Features (GLF), which results from theprojection of local surface patches into a common basis obtained from the GraphLaplacian eigenspace. We test the proposed approach in the BU-3DFE database interms of expressions and Action Units recognition. Our results confirm that theproposed GLF produces consistently higher recognition rates than thecurves-based approach, thanks to a more complete description of the surface,while requiring a lower computational complexity. We also show that the GLFoutperform the most popular alternative approach for spectral representation,Shape- DNA, which is based on the Laplace Beltrami Operator and cannot providea stable basis that guarantee that the extracted signatures for the differentpatches are directly comparable.
展开▼
