This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Object and Texture Classification Using Higher Order Statistics
July 1992 (vol. 14 no. 7)
pp. 733-750

The problem of the detection and classification of deterministic objects and random textures in a noisy scene is discussed. An energy detector is developed in the cumulant domain by exploiting the noise insensitivity of higher order statistics. An efficient implementation of this detector is described, using matched filtering. Its performance is analyzed using asymptotic distributions in a binary hypothesis-testing framework. The object and texture discriminant functions are minimum distance classifiers in the cumulant domain and can be efficiently implemented using a bank of matched filters. They are immune to additive Gaussian noise and insensitive to object shifts. Important extensions, which can handle object rotation and scaling, are also discussed. An alternative texture classifier is derived from a ML viewpoint and is statistically efficient at the expense of complexity. The application of these algorithms to the texture-modeling problem is indicated, and consistent parameter estimates are obtained.

[1] R. Bajcsy, "Computer description of textured surfaces," inProc. 3rd Int. Joint Conf. Artificial Intell., Aug. 1973, pp. 572-578.
[2] A. C. Bovik, M. Clark, and W. S. Geisler, "Multichannel texture analysis using localized spatial filters,"IEEE Trans. Patt. Anal. Machine Intell., vol. 12, pp. 55-73, Jan. 1990.
[3] D. R. Brillinger, "Fourier inference: Some methods for the analysis of array and non-Gaussian series data,"Water Resources Bull., vol. 21, no. 5, pp. 743-756, 1985.
[4] D. R. Brillinger and M. Rosenblatt, "Computation and interpretation ofkth-order spectra," inSpectral Analysis of Time Series(B. Harris, Ed.). New York: Wiley, 1967, pp. 189-232.
[5] T. Brodatz,Textures: A Photographic Album for Artists and Designers. New York: Dover, 1966.
[6] L. Capodiferro, R. Gusani, G. Jacoviti, and M. Vascotto, "A correlation based technique for shift, scale and rotation independent object identification," inProc. Int. Conf. Acoust. Speech Signal Processing(Dallas, TX), 1987, pp. 221-224.
[7] D. Casasent and D. Psaltis, "Position, rotation and scale invariant optical correlation,"Appl. Optics, vol. 15, pp. 1795-1799, Oct. 1976.
[8] R. Chellappa and S. Chatterjee, "Classification of textures using Gaussian Markov random fields,"IEEE Trans. Acoust. Speech Signal Processing, vol. ASSP-33, no. 4, pp. 959-963, 1985.
[9] C. H. Chen, "A study of texture classification using spectral features," inProc. Int. Conf. Patt. Recogn.(Munich, West Germany), 1982.
[10] H. J. Dohse, J. L. C. Sanz, and A. K. Jain, "Object classification and registration by radon transform based invariants," inProc. Int. Conf. Acoust. Speech Signal Processing(Dallas, TX), 1987, pp. 225-228.
[11] O. D. Faugeras, "Texture analysis and classification using a human visual system," inProc. Int. Joint Conf. Patt. Recogn.(Kyoto, Japan), Nov. 1978, pp. 549-559.
[12] O. D. Faugeras and W. K. Pratt, "Decorrelation methods of texture feature extraction,"IEEE Trans. Patt. Anal. Machine Intell., vol, PAMI- 2, pp. 323-332, July 1980.
[13] B. Friedlander and B. Porat, "Asymptotically optimal estimation of MA and ARMA parameters of non-Gaussian processes from higher-order moments,"IEEE Trans. Automat. Contr., vol. 35, pp. 27-35, Jan. 1990.
[14] G. B. Giannakis, "Signal reconstruction from multiple correlations: Frequency- and time-domain approaches,"J. Opt. Soc. Amer., vol. 6, pp. 682-697, May 1989.
[15] G. B. Giannakis, "Cumulants: A powerful tool in signal processing," inProc. IEEE, pp. 1333-1334, Sept. 1987.
[16] G. B. Giannakis, "On the identifiability of non-Gaussian ARMA models using cumulants,"IEEE Trans. Automat. Contr., vol. 35, pp. 18-26, Jan. 1990.
[17] G. B. Giannakis and J. M. Mendel, "Identification of non-minimum phase systems using higher-order statistics,"IEEE Trans. Acoust. Speech Signal Processing, vol. 37, pp. 360-377, 1989.
[18] G. B. Giannakis and A. Swami, "On estimating non-causal non-minimum phase ARMA models of non-Gaussian processes"IEEE Trans. Acoust. Speech Signal Processing, vol. 38, pp. 478-495, Mar. 1990.
[19] G. B. Giannakis and M. K. Tsatsanis, "Signal detection using matched filtering and higher-order statistics,"IEEE Trans. Acoust. Speech Signal Processing, vol. 38, pp. 1284-1296, July 1990.
[20] G. B. Giannakis and M. K. Tsatsanis, "A unifying maximum likelihood view of cumulant and polyspectral measures for nonGaussian signal estimation and classification,"IEEE Trans. Inform. Theory, vol. 38, pp. 386-406, Mar. 1992.
[21] T. E. Hall and G. B. Giannakis, "Texture model validation using higher-order statistics," inProc. Int. Conf. Acoust. Speech Signal Processing(Toronto), May 1991, pp. 2673-2676.
[22] R. M. Haralick, "Statistical and structural approach to texture,"Proc. IEEE, vol. 67, pp. 786-804, May 1979.
[23] M. J. Hinich, "Testing for Gaussianity and linearity of a stationary time-series,"J. Time Series Anal., pp. 169-176, 1982.
[24] Y. N. Hsu, H. H. Arsenault, and G. April, "Rotation invariant digital pattern recognition using circular harmonic expansion,"Appl. Opt., vol. 21, pp. 4012-4015, Nov. 1985.
[25] A. Jain,Fundamentals of Digital Image Processing. Englewood Cliffs, NJ: Prentice-Hall, 1989.
[26] R. L. Kashyap, "Univariate and multivariate random models for images,"Comput. Graphics Image Processing, vol. 12, pp. 257-270, 1980.
[27] R. L. Kashyap, "Characterization and estimation of two-dimensional ARMA model,"IEEE Trans. Inform. Theory, pp. 736-745, 1984.
[28] R. L. Kashyap, R. Chellapa, and A. Khotanzad, "Texture classification using features derived from random field models,"Patt. Recogn. Lett., pp. 43-50, 1982.
[29] R. L. Kashyap and A. Khotanzad, "A stochastic model based technique for texture segmentation," inProc. 7th Int. Conf. Patt. Recogn.(Montreal), July 1984, pp. 1202-1205.
[30] K. I. Laws, "Textured image segmentation," Ph.D. dissertation, Dept. Eng., Univ. of Southern California, Los Angeles, 1980.
[31] K. S. Lii and M. Rosenblatt, "Deconvolution and estimation of transfer function phase and coefficients for non-Gaussian linear processes,"Ann. Stat., pp. 1195-1208, 1982.
[32] A. W. Lohmann and B. Wirnitzer, "Triple correlations,"Proc. IEEE, vol. 72, pp. 889-901, 1984.
[33] O. Mitchell, C. Myers, and W. Boyne, "A max-min measure for image texture analysis,"IEEE Trans. Comput., vol. C-25, pp. 408-414, 1977.
[34] A. V. Oppenheim and R. W. Schafer,Digital Signal Processing. Englewood Cliffs, NJ: Prentice-Hall, 1975.
[35] B. Picinbono and P. Duvaut, "Optimal linear-quadratic systems for detection and estimation,"IEEE Trans. Inform. Theory, vol. 34, pp. 304-311, Mar. 1988.
[36] D. A. Pintsov, "Invariant pattern recognition, symmetry and radon transforms,"J. Opt. Soc. Amer., vol. 6, pp. 1544-1554, Oct. 1989.
[37] W. Pratt, O. Faugeras, and A. Gagalowicz, "Visual discrimination of stochastic texture-fields,"IEEE Trans. Syst. Man Cybern., vol. SMC-8, pp. 796-804, 1978.
[38] W. Pratt, O. Faugeras, and A. Gagalowicz, "Applications of stochastic texture field models to image processing,"Proc. IEEE, vol. 69, pp. 542-551, 1981.
[39] T. R. Reed and H. Wechsler, "Segmentation of textured images and Gestalt organization using spatial/spatial-frequency representations,"IEEE Trans. Patt Anal. Machine Intell., vol. 12, pp. 1-12, Jan. 1990.
[40] M. Rosenblatt,Stationary Sequences and Random Fields.--Boston: Birkhauser, 1985.
[41] M. Rosenblatt, and J. N. Van Ness, "Estimation of the bispectrum,"Ann. Math. Stat., pp. 1120-1136, 1965.
[42] A. Rosenfeld and M. Thurston, "Edge and curve detection for visual scene analysis,"IEEE Trans. Comput., vol. C-20, pp. 562-569, 1971.
[43] P. Ruiz and J. Lacoume, "Extraction of independent sources from correlated inputs," inProc. Workshop Higher Order Spectral Anal.(Vail, CO), June 1989, pp. 146-151.
[44] B. Sadler, and G. B. Giannakis, "Shift and rotation invariant object reconstruction using the bispectrum,"J. Opt. Soc. Amer. A, vol. 9, pp. 57-69, Jan. 1992.
[45] A. Swami, "System identification using cumulants," Ph.D. thesis, Univ. of Southern California, Los Angeles, Oct. 1988.
[46] A. Swami, G. B. Giannakis, and J. M. Mendel, "Linear modeling of multidimensional non-Gaussian processes using cumulants,"Multidimensional Syst. Signal Processing, vol. 1, pp. 11-37, Mar. 1990.
[47] C. W. Therrien, "An estimation-theoretic approach to terrain image segmentation,"Comput. Vision Graphics Image Processing, vol. 22, pp. 313-326, 1983.
[48] A. M. Tekalp, H. Kaufman, and J. W. Woods, "Identification of image and blur parameters for the restoration of non-causal filters,IEEE Trans. Acoust. Speech Signal Processing, vol. ASSP-34, pp. 963-972, 1986.
[49] M. K. Tsatsanis and G. B. Giannakis, "Object detection and classification using matched filtering and higher-order statistics," inProc. 6th Workshop Multidimensional Signal Processing(Monterey, CA), Sept. 1989, pp. 32-33.
[50] M. R. Turner, "Texture discrimination by Gabor functions,"Biolog. Cybern., vol. 55, pp. 71-82, 1986.
[51] H. L. Van Trees,Detection, Modulation, and Estimation Theory. New York: Wiley, 1969, vol. 1.
[52] H. Wechsler,Computational Vision. New York: Academic, 1990.
[53] J. Weszka, C. R. Dyer, and A. Rosenfeld, "A comparative study of texture measures for terrain classification,"IEEE Trans. Syst. Man Cybern., vol. SMC-6, pp. 269-285, Apr. 1976.
[54] J. M. Wozencraft and I. M. Jacobs,Principles of Communication Engineering. New York: Wiley, 1965.

Index Terms:
texture detection; object scaling; pattern recognition; texture classification; noisy scene; energy detector; higher order statistics; matched filtering; minimum distance classifiers; object rotation; parameter estimates; filtering and prediction theory; parameter estimation; pattern recognition; statistical analysis
Citation:
M.K. Tsatsanis, G.B. Giannakis, "Object and Texture Classification Using Higher Order Statistics," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 14, no. 7, pp. 733-750, July 1992, doi:10.1109/34.142910
Usage of this product signifies your acceptance of the Terms of Use.