| We show in this paper how Neural Networks can be used for Human Face Processing. In Part I, we show how Neural Networks can be viewed as a particular class of Statistical models. We introduce learning as an estimation problem, then describe Multi-Layer Perceptrons (MLP) and Radial Basis Function (RBF) networks, widely used Neural Networks which we will use in Part~II, for face processing. We further present Vapnik's framework for learning, show the capacity/generalization dilemma and discuss its implications for Neural Network training and model selection. Vapnik's ideas lead to a new interesting class of classifier, Support Vector Machines, presented in section 3. We then discuss the combination of models and give a formalism which allows to cooperatively train multi-modular Neural Networks architectures. Finally, we present a multi-modular architecture to perform "Segmentation-Recognition in the loop". In Part II, we show how the presented models can be applied to build an efficient face localization and identification system. The face images are detected by scanning the scene with a retina feeding a hierarchical coarse-to-fine classifier. Detections are then identified in a small family of known persons. |
Récupérer au format
