Texture segmentation through eigen-analysis of the Pseudo-Wigner distribution

Abstract In this paper we propose a new method for texture segmentation based on the use of texture feature detectors derived from a decorrelation procedure of a modified version of a Pseudo-Wigner distribution (PWD). The decorrelation procedure is accomplished by a cascade recursive least squared (CRLS) principal component (PC) neural network. The goal is to obtain a more efficient analysis of images by combining the advantages of using a high-resolution joint representation given by the PWD with an effective adaptive principal component analysis (PCA) through the use of feedforward neural networks.

[1]  Andrzej Cichocki,et al.  Adaptive learning algorithm for principal component analysis with partial data , 1996 .

[2]  J. Mayer,et al.  On the Quantum Correction for Thermodynamic Equilibrium , 1947 .

[3]  Boualem Boashash,et al.  Methods of signal classification using the images produced by the Wigner-Ville distribution , 1991, Pattern Recognit. Lett..

[4]  Robert Goutte,et al.  On the use of two-dimensional Wigner-Ville distribution for texture segmentation , 1993, Signal Process..

[5]  H. Suzuki,et al.  A method of two‐dimensional spectral analysis using the wigner distribution , 1992 .

[6]  E. Oja.  Simplified neuron model as a principal component analyzer , 1982, Journal of mathematical biology.

[7]  Harry Wechsler,et al.  Segmentation of Textured Images and Gestalt Organization Using Spatial/Spatial-Frequency Representations , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Ljubisa Stankovic,et al.  Auto-term representation by the reduced interference distributions: a procedure for kernel design , 1996, IEEE Trans. Signal Process..

[9]  Josef Bigün,et al.  N-folded Symmetries by Complex Moments in Gabor Space and their Application to Unsupervised Texture Segmentation , 1994, IEEE Trans. Pattern Anal. Mach. Intell..

[10]  Dennis Gabor,et al.  Theory of communication , 1946 .

[11]  S. Hahn,et al.  Multidimensional complex signals with single-orthant spectra , 1992, Proc. IEEE.

[12]  Josef Bigün,et al.  Hierarchical image segmentation by multi-dimensional clustering and orientation-adaptive boundary refinement , 1995, Pattern Recognit..

[13]  Simon Haykin,et al.  Neural networks expand SP's horizons , 1996, IEEE Signal Process. Mag..

[14]  P. Flandrin,et al.  Detection of changes of signal structure by using the Wigner-Ville spectrum , 1985 .

[15]  T. Claasen,et al.  THE WIGNER DISTRIBUTION - A TOOL FOR TIME-FREQUENCY SIGNAL ANALYSIS , 1980 .

[16]  Bernd Fritzke.  Vector Quantization with a Growing and Splitting Elastic Net , 1993 .

[17]  C Gorecki.  Surface classification by an optoelectronic implementation of the Karhunen-Loève expansion. , 1991, Applied optics.

[18]  Anil K. Jain,et al.  Learning Texture Discrimination Masks , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[19]  Douglas L. Jones,et al.  A resolution comparison of several time-frequency representations , 1992, IEEE Trans. Signal Process..

[20]  Mahmood R. Azimi-Sadjadi,et al.  Principal component extraction using recursive least squares learning , 1995, IEEE Trans. Neural Networks.

[21]  Simon Haykin,et al.  Wigner-Ville distribution: an important functional block for radar target detection in clutter , 1994, Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers.

[22]  Gabriel Cristóbal,et al.  High resolution spectral analysis of images using the pseudo-Wigner distribution , 1998, IEEE Trans. Signal Process..

[23]  P. Flandrin,et al.  Méthodes temps-fréquence , 1992 .

[24]  Terence D. Sanger,et al.  Optimal unsupervised learning in a single-layer linear feedforward neural network , 1989, Neural Networks.