@MISC\{IMM2002-0987,
    author       = "K. B. Hilger and M. B. Stegmann and R. Larsen",
    title        = "A Noise Robust Statistical Texture Model",
    year         = "2002",
    publisher    = "",
    address      = "",
    url          = "http://www2.compute.dtu.dk/pubdb/pubs/987-full.html",
    abstract     = "The study presents a noise robust low dimensional representation of texture variation present in a training set. The conventional analysis of training textures in the Active Appearance Models (AAMs) segmentation frame-work is extended by the new representation. This is accomplished by augmenting the model with an estimate of the covariance of the noise present in the training data. A compact model is obtained maximizing the signal-to-noise ratio (SNR), thus favouring subspaces rich on signal, and low on noise. The extended method is evaluated on a set of left cardiac ventricles obtained using magnetic resonance imaging (MRI)."
}