Browsing by Author "Musa, MEM"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Article
    Citation - WoS: 4
    Citation - Scopus: 5
    Almost Autonomous Training of Mixtures of Principal Component Analyzers
    (Elsevier Science Bv, 2004) Musa, MEM; de Ridder, D; Duin, RPW; Atalay, V
    In recent years, a number of mixtures of local PCA models have been proposed. Most of these models require the user to set the number of submodels (local models) in the mixture and the dimensionality of the submodels (i.e., number of PC's) as well. To make the model free of these parameters, we propose a greedy expectation-maximization algorithm to find a suboptimal number of submodels. For a given retained variance ratio, the proposed algorithm estimates for each submodel the dimensionality that retains this given variability ratio. We test the proposed method on two different classification problems: handwritten digit recognition and 2-class ionosphere data classification. The results show that the proposed method has a good performance. (C) 2004 Elsevier B.V. All rights reserved.
  • Thumbnail Image
    Conference Object
    Citation - WoS: 1
    Citation - Scopus: 2
    Texture Segmentation Using the Mixtures of Principal Component Analyzers
    (Springer-verlag Berlin, 2003) Musa, MEM; Duin, RPW; de Ridder, D; Atalay, V
    The problem of segmenting an image into several modalities representing different textures can be modelled using Gaussian mixtures. Moreover, texture image patches when translated, rotated or scaled lie in low dimensional subspaces of the high-dimensional space spanned by the grey values. These two aspects make the mixture of local subspace models worth consideration for segmenting this type of images. In recent years a number of mixtures of local PCA models have been proposed. Most of these models require the user to set the number of subspaces and subspace dimensionalities. To make the model autonomous, we propose a greedy EM algorithm to find a suboptimal number of subspaces, besides using a global retained variance ratio to estimate for each subspace the dimensionality that retains the given variability ratio. We provide experimental results for testing the proposed method on texture segmentation.