Browsing by Author "Atalay, Volkan"

Now showing 1 - 5 of 5

Citation Count: Musa, MEM; de Ridder, D.; Duin, RPW; Atalay, V., "Almost autonomous training of mixtures of principal component analyzers" Pattern Recognition Letters, Vol.25, No.9, pp.1085-1095, (2004).
Almost autonomous training of mixtures of principal component analyzers
(Elsevier Science BV, 2004) Musa, Mohamed E. M.; Ridder, Dick de; Duin, Robert P. W.; Atalay, Volkan
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.
Citation Count: Hassanpour, R.,; Atalay, V., "Delaunay Triangulation Based 3D Human Face Modeling From Uncalibrated İmages", IEEE Computer Society Conference On Computer Vision and Pattern Recognition Workshops, (2004).
Delaunay Triangulation Based 3D Human Face Modeling From Uncalibrated İmages
(IEEE Computer Society, 2004) Hassanpour, Reza; Atalay, Volkan; 48646
In this paper, we describe an algorithm for generating three dimensional models of human faces from uncalibrated images. Input images are taken by a camera generally with a small rotation around a single axis which may cause degenerate solutions during auto-calibration. We describe a solution to this problem by a priori assumptions on the camera. To generate a specific person's head, a generic human head model is deformed according to the 3D coordinates of points obtained by reconstructing the scene using images calibrated with our algorithm. The deformation process is based on a physical based massless spring model and it requires local re-triangulation in the area with high curvatures. This is achieved by locally applying Delaunay traingulation method. However, there may occur degeneracies in Delaunay triangulation such as encroaching of edges. We describe an algorithm for removing the degeneracies during triangulation by modifying the definition of the Delaunay cavity. This algorithm has also the effect of preserving the curavature in the face area. We have compared the models generated with our algorithm with the models obtained using cyberscanners. The RMS geometric error in these comparisons are less than 1.8 x 10-2.
Citation Count: Hassanpour, R., Atalay, V., (2006). Experimental study on the sensitivity of autocalibration to projective camera model parameters. Optical Engineering, 45(4). http://dx.doi.org/10.1117/1.2189292
Experimental study on the sensitivity of autocalibration to projective camera model parameters
(Spie-Int Society Optical Engineering, 2006) Hassanpour, Reza; Atalay, Volkan; 48646
Existing methods of 3-D object modeling and recovering 3-D data from uncalibrated 2-D images are subject to errors introduced by assumptions about camera parameters and mismatches in finding point pairs in the images. In this study, we experimentally evaluate the effect of each of these assumptions together with the inaccuracy in the measurements in the images. Sensitivity of reconstruction errors to inaccuracies in the estimation of camera parameters and mismatches due to noise in input data is measured using a linear and two nonlinear autocalibration methods for a projective camera. Our experimental results show that some assumptions such as a vanishing skew can be safely made; however, other parameters such as principal point location are quite sensitive to wrong assumptions. (c) 2006 Society of Photo-Optical Instrumentation Engineers. Keywords
Citation Count: Hassanpour, Reza; Atalay, Volkan. "Head Modeling with Camera Auto-calibration and Deformation", 2002.
Head Modeling with Camera Auto-calibration and Deformation
(2002) Hassanpour, Reza; Atalay, Volkan
A 3D head modeling method from a sequence of 2D images is described. The views from which the input images are acquired are not calibrated. Therefore, an auto-calibration method for a sequence of images with small rotations and translation is developed. For this purpose, we have modified an already existing auto-calibration algorithm to incorporate known aspect ratio and skew values to make it applicable for small rotation around a single axis. We apply this auto-calibration technique to head (face) modeling. Three dimensional positions of known facial features computed from two dimensional images are used to deform a generic head model by using a spring based energy minimization method.
Citation Count: Musa, Mohamed E.M...et al (2003). "Texture segmentation using the mixtures of principal component analyzers", Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 2869, pp. 505-512.
Texture segmentation using the mixtures of principal component analyzers
(2003) Musa, Mohamed E.M.; Duin, Robert P.W.; De Ridder, Dick; Atalay, Volkan
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.