Face Recognition by Elastic Bunch Graph Matching*†
Laurenz Wiskott, Jean‐Marc Fellous, Nobert Krüger, Christoph von der Malsburg
Abstract
Classification systems differ vastly in terms of the nature and origin of their knowledge about image variations. Systems in Artificial Intelligence and Computer Vision often stress specific designer-provided structures, for instance explicit models of three-dimensional objects or of the image-generation process, whereas Neural Network models tend to stress absorption of structure from examples with the help of statistical estimation techniques. Gabor wavelets are biologically motivated convolution kernels in the shape of plane waves restricted by a Gaussian envelope function. Some face recognition systems are based on user-defined face-specific features. Yuille, for example, represented eyes by a circle within an almond-shape and defined an energy function to optimize a total of 9 model parameters for matching it to an image. The manual definition of appropriate grid structures and the semiautonomous process of bunch graph acquisition will have to be replaced by a fully autonomous process.