In this paper, an idea for the combination of global and local view models is presented. These two types of models have proved their capabilities independently. Some combination were proposed, using global view models for local analysis, and local view models to synthetize local results. An opposite approach is proposed here: local view models are used as a normalization tool, while global view models are used for the recognition of the normalized image. The use of local view models for normalization is justified by their capability to perform a non-linear normalization according to the image information. We propose Markov models as local view models, and Neural Network as global view models. Using Markov models for the normalization increases results up to 3% better than a classical linear normalization. Global results are improved, performing 2% better than the Markov model itself. The extension of the system to an analytic approach is discussed.