Abstract

We present in detail the recorded results of the modified-hybrid optical neural network (M-HONN) filter during a full series of tests to examine its robustness and overall performance for object recognition tasks. We test the M-HONN filter for its detectability and peak sharpness with within-class distortion of the input object, its discrimination ability between an in-class and out-of-class object, and its performance with cluttered images of the true-class object. The M-HONN filter is found to exhibit good detectability, an ability to maintain its correlation-peak sharpness throughout the recorded tests, good discrimination ability, and an ability to detect the true-class object within cluttered input images. Additionally we observe the M-HONN filter’s performance within the tests in comparison with the constrained-hybrid optical neural network filter for the first three series of tests and the synthetic discriminant function-maximum average correlation height filter for the fourth set of tests.

© 2008 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. C. G. Looney, Pattern Recognition Using Neural Networks-Theory and Algorithms for Engineers and Scientists (Oxford University Press, 1997).
  2. S. Haykin, Neural Networks--A Comprehensive Foundation (Prentice Hall, 1994).
  3. R. Beale and T. Jackson, Neural Computing: An Introduction (Institute of Physics Publishing, 1990).
    [CrossRef]
  4. B. V. K. Vijaya Kumar, “Tutorial survey of composite filter designs for optical correlators,” Appl. Opt. 31, 4773-4801 (1992).
    [CrossRef]
  5. I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).
  6. I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
    [CrossRef]
  7. I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
    [CrossRef]
  8. L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
    [CrossRef]
  9. L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
    [CrossRef]
  10. D. Casasent, “Unified synthetic discriminant function computational formulation,” Appl. Opt. 23, 1620-1627 (1984).
    [CrossRef] [PubMed]
  11. H. J. Caulfield, “Linear combinations of filters for character recognition: a unified treatment,” Appl. Opt. 19, 3877-3878(1980).
    [CrossRef] [PubMed]
  12. H. J. Caulfield and W. T. Maloney, “Improved discrimination in optical character recognition,” Appl. Opt. 08, 2354-2356(1969).
    [CrossRef]
  13. L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
    [CrossRef]
  14. L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
    [CrossRef]
  15. B. V. K. Vijaya Kumar and L. Hassebrook, “Performance measures for correlation filters,” Appl. Opt. 29, 2997-3006(1990).
    [CrossRef]
  16. P. Refregier, “Filter design for optical pattern recognition: multicriteria optimisation approach,” Opt. Lett. 15, 854-856(1990).
    [CrossRef] [PubMed]
  17. P. Refregier, “Optimal trade-off filters for noise robustness, sharpness of the correlation peak and Horner efficiency,” Opt. Lett. 16, 829-831 (1991).
    [CrossRef] [PubMed]
  18. H. Demuth and M. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide for MATLAB Version 6.5 (MathWorks, Inc., 2002), http://www.mathworks.com.
  19. M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).
  20. D. Nguyen and B. Widrow, “The truck backer-upper: an example of self-learning in neural networks,” in Proceedings of the International Joint Conference on Neural Networks (IEEE, 1989), Vol. 2, pp. 357-363.
    [CrossRef]
  21. D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of the International Joint Conference on Neural Networks, (IEEE, 1990), Vol. 3, pp. 21-26.
    [CrossRef]
  22. A. Talukder and D. Casasent, “Non-linear features for product inspection,” Proc. SPIE 3715, 32-43 (1999).
    [CrossRef]
  23. D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
    [CrossRef]
  24. I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
    [CrossRef]
  25. A. Mahalanobis, B. V. K. Vijaya Kumar, S. Song, S. R. F. Sims, and J. F. Epperson, “Unconstrained correlation filters,” Appl. Opt. 33, 3751-3759 (1994).
    [CrossRef] [PubMed]
  26. A. Vander Lugt, “Signal detection by complex spatial filtering,” IEEE Trans. Inf. Theory 10, 139-145 (1964).
    [CrossRef]
  27. T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
    [CrossRef]
  28. H. Zhou and T. H. Chao, “MACH filter synthesising for detecting targets in cluttered environment for gray-scale optical correlator,” Proc. SPIE 3715, 394-398 (1999).
    [CrossRef]
  29. S. Goyal, N. K. Nishchal, V. K. Beri, and A. K. Gupta, “Wavelet-modified maximum average correlation height filter for rotation invariance that uses chirp encoding in a hybrid digital-optical correlator,” Appl. Opt. 45, 4850-4857 (2006).
    [CrossRef] [PubMed]
  30. E. Stamos, Similarity suppression filter design algorithm, in “Algorithms for designing filters for optical pattern recognition,” D. Phil. Thesis (University College London, 2001), pp. 77-78.
  31. B. V. K. Vijaya Kumar, “Minimum variance synthetic discriminant functions,” J. Opt. Soc. Am. A 3, 1579-1584 (1986).
    [CrossRef]
  32. I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.
  33. A. Mahalanobis and B. V. K. Vijaya Kumar, “Optimality of the maximum average correlation height filter for detection of targets in noise,” Opt. Eng. 36, 2642-2648 (1997).
    [CrossRef]

2006

2004

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
[CrossRef]

2003

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

2002

H. Demuth and M. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide for MATLAB Version 6.5 (MathWorks, Inc., 2002), http://www.mathworks.com.

I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).

2001

E. Stamos, Similarity suppression filter design algorithm, in “Algorithms for designing filters for optical pattern recognition,” D. Phil. Thesis (University College London, 2001), pp. 77-78.

1999

A. Talukder and D. Casasent, “Non-linear features for product inspection,” Proc. SPIE 3715, 32-43 (1999).
[CrossRef]

H. Zhou and T. H. Chao, “MACH filter synthesising for detecting targets in cluttered environment for gray-scale optical correlator,” Proc. SPIE 3715, 394-398 (1999).
[CrossRef]

1998

T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
[CrossRef]

D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
[CrossRef]

1997

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
[CrossRef]

A. Mahalanobis and B. V. K. Vijaya Kumar, “Optimality of the maximum average correlation height filter for detection of targets in noise,” Opt. Eng. 36, 2642-2648 (1997).
[CrossRef]

C. G. Looney, Pattern Recognition Using Neural Networks-Theory and Algorithms for Engineers and Scientists (Oxford University Press, 1997).

1996

M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).

1994

1992

1991

1990

P. Refregier, “Filter design for optical pattern recognition: multicriteria optimisation approach,” Opt. Lett. 15, 854-856(1990).
[CrossRef] [PubMed]

B. V. K. Vijaya Kumar and L. Hassebrook, “Performance measures for correlation filters,” Appl. Opt. 29, 2997-3006(1990).
[CrossRef]

R. Beale and T. Jackson, Neural Computing: An Introduction (Institute of Physics Publishing, 1990).
[CrossRef]

D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of the International Joint Conference on Neural Networks, (IEEE, 1990), Vol. 3, pp. 21-26.
[CrossRef]

1989

D. Nguyen and B. Widrow, “The truck backer-upper: an example of self-learning in neural networks,” in Proceedings of the International Joint Conference on Neural Networks (IEEE, 1989), Vol. 2, pp. 357-363.
[CrossRef]

1986

1984

1980

1969

1964

A. Vander Lugt, “Signal detection by complex spatial filtering,” IEEE Trans. Inf. Theory 10, 139-145 (1964).
[CrossRef]

Beale, M.

H. Demuth and M. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide for MATLAB Version 6.5 (MathWorks, Inc., 2002), http://www.mathworks.com.

Beale, M. H.

M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).

Beale, R.

R. Beale and T. Jackson, Neural Computing: An Introduction (Institute of Physics Publishing, 1990).
[CrossRef]

Beri, V. K.

Birch, P.

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

Casasent, D.

A. Talukder and D. Casasent, “Non-linear features for product inspection,” Proc. SPIE 3715, 32-43 (1999).
[CrossRef]

D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
[CrossRef]

D. Casasent, “Unified synthetic discriminant function computational formulation,” Appl. Opt. 23, 1620-1627 (1984).
[CrossRef] [PubMed]

Caulfield, H. J.

Chao, T. H.

H. Zhou and T. H. Chao, “MACH filter synthesising for detecting targets in cluttered environment for gray-scale optical correlator,” Proc. SPIE 3715, 394-398 (1999).
[CrossRef]

T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
[CrossRef]

Chatwin, C.

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

Chatwin, C. R.

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
[CrossRef]

I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
[CrossRef]

I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.

Demuth, H.

H. Demuth and M. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide for MATLAB Version 6.5 (MathWorks, Inc., 2002), http://www.mathworks.com.

Demuth, H. B.

M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).

Epperson, J. F.

Goyal, S.

Gupta, A. K.

Hagan, M. T.

M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).

Hassebrook, L.

Haykin, S.

S. Haykin, Neural Networks--A Comprehensive Foundation (Prentice Hall, 1994).

Jackson, T.

R. Beale and T. Jackson, Neural Computing: An Introduction (Institute of Physics Publishing, 1990).
[CrossRef]

Jamal-Aldin, L. S.

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
[CrossRef]

Kypraios, I.

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
[CrossRef]

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).

I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.

Lei, P.

I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.

Looney, C. G.

C. G. Looney, Pattern Recognition Using Neural Networks-Theory and Algorithms for Engineers and Scientists (Oxford University Press, 1997).

Mahalanobis, A.

A. Mahalanobis and B. V. K. Vijaya Kumar, “Optimality of the maximum average correlation height filter for detection of targets in noise,” Opt. Eng. 36, 2642-2648 (1997).
[CrossRef]

A. Mahalanobis, B. V. K. Vijaya Kumar, S. Song, S. R. F. Sims, and J. F. Epperson, “Unconstrained correlation filters,” Appl. Opt. 33, 3751-3759 (1994).
[CrossRef] [PubMed]

Maloney, W. T.

Neiberg, L. M.

D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
[CrossRef]

Nguyen, D.

D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of the International Joint Conference on Neural Networks, (IEEE, 1990), Vol. 3, pp. 21-26.
[CrossRef]

D. Nguyen and B. Widrow, “The truck backer-upper: an example of self-learning in neural networks,” in Proceedings of the International Joint Conference on Neural Networks (IEEE, 1989), Vol. 2, pp. 357-363.
[CrossRef]

Nishchal, N. K.

Park, Y.

T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
[CrossRef]

Refregier, P.

Reyes, G.

T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
[CrossRef]

Sims, S. R. F.

Sipe, M. A.

D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
[CrossRef]

Song, S.

Stamos, E.

E. Stamos, Similarity suppression filter design algorithm, in “Algorithms for designing filters for optical pattern recognition,” D. Phil. Thesis (University College London, 2001), pp. 77-78.

Talukder, A.

A. Talukder and D. Casasent, “Non-linear features for product inspection,” Proc. SPIE 3715, 32-43 (1999).
[CrossRef]

Vander Lugt, A.

A. Vander Lugt, “Signal detection by complex spatial filtering,” IEEE Trans. Inf. Theory 10, 139-145 (1964).
[CrossRef]

Vijaya Kumar, B. V. K.

Widrow, B.

D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of the International Joint Conference on Neural Networks, (IEEE, 1990), Vol. 3, pp. 21-26.
[CrossRef]

D. Nguyen and B. Widrow, “The truck backer-upper: an example of self-learning in neural networks,” in Proceedings of the International Joint Conference on Neural Networks (IEEE, 1989), Vol. 2, pp. 357-363.
[CrossRef]

Young, R.

I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).

Young, R. C. D

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

Young, R. C. D.

I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
[CrossRef]

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
[CrossRef]

I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.

Zhou, H.

H. Zhou and T. H. Chao, “MACH filter synthesising for detecting targets in cluttered environment for gray-scale optical correlator,” Proc. SPIE 3715, 394-398 (1999).
[CrossRef]

Appl. Opt.

H. J. Caulfield, “Linear combinations of filters for character recognition: a unified treatment,” Appl. Opt. 19, 3877-3878(1980).
[CrossRef] [PubMed]

D. Casasent, “Unified synthetic discriminant function computational formulation,” Appl. Opt. 23, 1620-1627 (1984).
[CrossRef] [PubMed]

B. V. K. Vijaya Kumar and L. Hassebrook, “Performance measures for correlation filters,” Appl. Opt. 29, 2997-3006(1990).
[CrossRef]

B. V. K. Vijaya Kumar, “Tutorial survey of composite filter designs for optical correlators,” Appl. Opt. 31, 4773-4801 (1992).
[CrossRef]

A. Mahalanobis, B. V. K. Vijaya Kumar, S. Song, S. R. F. Sims, and J. F. Epperson, “Unconstrained correlation filters,” Appl. Opt. 33, 3751-3759 (1994).
[CrossRef] [PubMed]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Application of non-linearity to wavelet-transformed images to improve correlation filter performance,” Appl. Opt. 36, 9212-9224 (1997).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Synthetic discriminant function filter employing nonlinear space-domain preprocessing on bandpass-filtered images,” Appl. Opt. 37, 2051-2062 (1998).
[CrossRef]

S. Goyal, N. K. Nishchal, V. K. Beri, and A. K. Gupta, “Wavelet-modified maximum average correlation height filter for rotation invariance that uses chirp encoding in a hybrid digital-optical correlator,” Appl. Opt. 45, 4850-4857 (2006).
[CrossRef] [PubMed]

H. J. Caulfield and W. T. Maloney, “Improved discrimination in optical character recognition,” Appl. Opt. 08, 2354-2356(1969).
[CrossRef]

Asian J. Phys.

I. Kypraios, R. Young, and C. R. Chatwin, “An investigation of the non-linear properties of correlation filter synthesis and neural network design,” Asian J. Phys. 11, 313-344 (2002).

IEEE Trans. Inf. Theory

A. Vander Lugt, “Signal detection by complex spatial filtering,” IEEE Trans. Inf. Theory 10, 139-145 (1964).
[CrossRef]

J. Opt. Soc. Am. A

Opt. Eng.

A. Mahalanobis and B. V. K. Vijaya Kumar, “Optimality of the maximum average correlation height filter for detection of targets in noise,” Opt. Eng. 36, 2642-2648 (1997).
[CrossRef]

D. Casasent, L. M. Neiberg, and M. A. Sipe, “Feature space trajectory distorted object representation for classification and pose estimation,” Opt. Eng. 37, 914-923 (1998).
[CrossRef]

I. Kypraios, R. C. D. Young, P. Birch, and C. R. Chatwin, “Object recognition within cluttered scenes employing a hybrid optical neural network (HONN) filter,” Opt. Eng. 43, 1839-1850 (2004).
[CrossRef]

Opt. Lett.

Pattern Recogn.

I. Kypraios, P. Lei, R. C. D. Young, and C. R. Chatwin, “Object recognition within cluttered scenes employing the modified-hybrid optical neural network filter,” submitted to Pattern Recogn.

Proc. SPIE

A. Talukder and D. Casasent, “Non-linear features for product inspection,” Proc. SPIE 3715, 32-43 (1999).
[CrossRef]

I. Kypraios, R. C. D Young, P. Birch, and C. Chatwin, “A non-linear training set superposition filter derived by neural network training methods for implementation in a shift invariant optical correlator,” Proc. SPIE 5106, 84-95 (2003).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “In-class distortion tolerance, out-of-class discrimination and clutter resistance of correlation filters that employ a space domain non-linearity applied to wavelet filtered input images,” Proc. SPIE 3386, 111-122 (1998).
[CrossRef]

L. S. Jamal-Aldin, R. C. D. Young, and C. R. Chatwin, “Nonlinear preprocessing operation for enhancing correlator filter performance in clutter,” Proc. SPIE 3490, 182-186 (1998).
[CrossRef]

I. Kypraios, R. C. D. Young, and C. R. Chatwin, “Performance assessment of unconstrained hybrid optical neural network (U-HONN) filter for object recognition tasks in clutter,” Proc. SPIE 5437, 51-62 (2004).
[CrossRef]

T. H. Chao, G. Reyes, and Y. Park, “Grayscale optical correlator,” Proc. SPIE 3386, 60-64 (1998).
[CrossRef]

H. Zhou and T. H. Chao, “MACH filter synthesising for detecting targets in cluttered environment for gray-scale optical correlator,” Proc. SPIE 3715, 394-398 (1999).
[CrossRef]

Other

E. Stamos, Similarity suppression filter design algorithm, in “Algorithms for designing filters for optical pattern recognition,” D. Phil. Thesis (University College London, 2001), pp. 77-78.

C. G. Looney, Pattern Recognition Using Neural Networks-Theory and Algorithms for Engineers and Scientists (Oxford University Press, 1997).

S. Haykin, Neural Networks--A Comprehensive Foundation (Prentice Hall, 1994).

R. Beale and T. Jackson, Neural Computing: An Introduction (Institute of Physics Publishing, 1990).
[CrossRef]

H. Demuth and M. Beale, Neural Network Toolbox for Use with MATLAB: User's Guide for MATLAB Version 6.5 (MathWorks, Inc., 2002), http://www.mathworks.com.

M. T. Hagan, H. B. Demuth, and M. H. Beale, Neural Network Design (PWS Publishing, 1996).

D. Nguyen and B. Widrow, “The truck backer-upper: an example of self-learning in neural networks,” in Proceedings of the International Joint Conference on Neural Networks (IEEE, 1989), Vol. 2, pp. 357-363.
[CrossRef]

D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of the International Joint Conference on Neural Networks, (IEEE, 1990), Vol. 3, pp. 21-26.
[CrossRef]

Cited By

OSA participates in CrossRef's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (8)

Fig. 1
Fig. 1

Architecture of the selected artificial NNET block of the HONN filter.

Fig. 2
Fig. 2

Applying a nonlinear optical mask in the input of the HONN filter. Here the mask is built by constant weight values equal to a chosen image of the NNET instead of multiplying each image with the corresponding weight connections of the hidden layer specific to that image. The M-HONN filter is based on the ability of the NNET to generalize with a nonlinear interpolation between the referenced and nonreferenced images presented to the NNET.

Fig. 3
Fig. 3

(a) Jaguar model, (b) Mazda model, and (c)–(f) background car park scenes—part of the third data set used in the experiments.

Fig. 4
Fig. 4

(a) Composite image of the basic SDF filter for the training set over the range of [ 20 ° 40 ° 60 ° 70 ° ] of the car images. (b) The composite image of the SDF-MACH filter for the training set over the same range of the car images. (c) The composite image of the M-HONN filter over the same range of the training images for T true = + 80 and T false = 80 . (d) The composite image of the M-HONN filter over the same range of the training images for T true = + 280 and T false = 280 . All the images of the training set used are gray-scale of size [ 256 × 256 ] .

Fig. 5
Fig. 5

(a) and (b) Correlation plane isometric and gray-level image, respectively, of the M-HONN filter for the training set over the orientation range of [ 20 ° 40 ° 60 ° 70 ° ] of the car images for the test image of 80 orientation T true = + 80 and T false = 80 . (c) and (d) The correlation plane isometric and the gray-level image, respectively, of the M-HONN filter for the same training set and the same test image, but T true = + 280 and T false = 280 .

Fig. 6
Fig. 6

(a) Correlation-peak height versus the angles of view over the range of [ 20 ° 70 ° ] for each input image for the M-HONN filter. The training set consisted of images at increments of 20 . (We tested the M-HONN filter with the object’s intermediate car poses over the same range at 10 increments.) (b) The nonnormalized PCE values of the test images at 10 increments versus the angles of view over the range [ 20 ° 70 ° ] for the M-HONN filter.

Fig. 7
Fig. 7

(a) Reference angle, Θ 0 , and the two in-class training images at the angles Θ 1 and Θ 2 . The test image is at the bisector at angle Θ 2 . (b) The correlation-peak height for each input image over a range of Θ ^ 3 [ 5 ° , 40 ° ] for the M-HONN filter. It is apparent the filter has good performance in recognizing all the intermediate car poses of the test set.

Fig. 8
Fig. 8

(a) Jaguar model inserted in a background scene of a car park and (b) the background scene used.

Tables (2)

Tables Icon

Table 1 Discrimination Ability of the M-HONN Filter for the In-Class Training Jaguar Model Image at 40 ° and the Out-of-Class Nontraining Mazda Model Image at the Same Angle

Tables Icon

Table 2 Target-to-Clutter Ratio

Equations (21)

Equations on this page are rendered with MathJax. Learn more.

h ( k , l ) = i = 1 N a i x i ( k , l ) ,
a = R 1 c ,
net p κ = p ( w ι κ o p κ + b ι ) ,
net x i = ι = l m × n w ι x i s ι x i ,
n e t x N = ι = 1 m × n w ι x N s ι x N .
N iw = N × [ m × n ] ,
N = 1 , 2 , 3 , , i ,
N 1 = N .
N l w = N × N opn ,
N b = N ,
Δ w ( i , i + 1 ) = μ × Δ w ( i 1 , i ) + α × μ × Δ P f Δ w ( i + 1 , i ) ,
Δ b ( i , i + 1 ) = μ × Δ b ( i 1 , i ) + α × μ × Δ P f Δ b ( i + 1 , i ) ,
α = { α = α + ϵ if     Δ P f < 0 α = no     change if     0 < Δ P f and Δ P f > max ( P f ) α = α ϵ if     Δ P f > max ( P f ) } ,
U-HONN = i = 1 N N S i ( m , n ) .
C-HONN = i = 1 N N a i · S i ( m , n ) .
Γ c = W x c · L x c = [ w 11 x c w 12 x c w 1 n 1 x c w 1 n x c w 21 x c w 22 x c w 2 n 1 x c w 2 n x c w m 1 x c w m 2 x c w m n 1 x c w m n x c ] · [ l 11 x c l l q x c l 21 x c l 2 q x c l n 1 x c l n q x c ] ,
S i = 1 N = Γ c · X i = 1 N ( m , n ) .
M-HONN = i = 1 N N a i · S i ( m , n ) .
N iw = 10 × [ 256 × 256 ] = 10 × 65 , 536 = 655 , 360.
Δ T = | T true T false | ,
Γ 60 ° = W x 60 ° · L x 60 ° = [ w 11 x 60 ° w 12 x 60 ° w 1 n 1 x 60 ° w 1 n x 60 ° w 21 x 60 ° w 22 x 60 ° w 2 n 1 x 60 ° w 2 n x 60 ° w m 1 x 60 ° w m 2 x 60 ° w m n 1 x 60 ° w m n x 60 ° ] · [ l 11 x 60 ° l 1 q x 60 ° l 21 x 60 ° l 2 q x 60 ° l n 1 x 60 ° l n q x 60 ° ] ,

Metrics