Abstract

Image enhancement has been shown to improve face recognition by visually impaired observers. We conducted three experiments in an effort to refine our understanding of the parameters leading to this effect. In experiment 1 we found that the band of spatial frequencies between 4 and 8 cycles/face is critical for face recognition. In experiment 2 we found that enhancement of these frequencies and the resulting image distortion actually reduced recognition performance for normal observers. Since the degradation of performance by low vision is larger than the effect of distortion, the enhancement that reduces performance for normal observers may still be beneficial for the visually impaired observer. Experiment 3 found that patients tend to prefer images enhanced at frequencies higher than the critical frequencies found in experiment 1. Such individually selected enhancement did not improve recognition in comparison with uniformly applied enhancement. The lack of an enhancement effect may be due to the small variability in enhancement frequencies selected by our subject population.

© 1994 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
    [PubMed]
  2. E. Peli, T. Peli, “Image enhancement for the visually impaired,” Opt. Eng. 23, 47–51 (1984).
    [CrossRef]
  3. E. Peli, “Limitations of image enhancement for the visually impaired,” J. Opt. Vis. Sci. 69, 15–24 (1992).
    [CrossRef]
  4. A. P. Ginsburg, “Specifying relevant spatial information for image evaluation and display design: an explanation of how we see certain objects,” Proc. Soc. Inf. Disp. 21, 219–227 (1980).
  5. A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
    [CrossRef] [PubMed]
  6. G. S. Rubin, K. Siegel, “Recognition of low pass filtered faces and letters,” Invest. Ophthalmol. Vis. Sci. Suppl. 25, 28 (1984).
  7. J. Sergent, “Microgenesis of face perception,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, R. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 17–33.
    [CrossRef]
  8. L. O. Harvey, G. P. Sinclair, “On the quality of visual imagery,” Invest, Ophthalmol. Vis. Sci. Suppl. 26, 281 (1985).
  9. M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).
  10. E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.
  11. R. A. Schuchard, G. S. Rubin, “Face identification of band pass filtered faces by low vision observers,” Invest. Ophthalmol. Vis. Sci. Suppl. 30, 396 (1989).
  12. T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
    [CrossRef]
  13. E. Peli, “Display nonlinearity in digital image processing for visual communication,” Opt. Eng. 31, 2374–2382 (1992).
    [CrossRef]
  14. E. Peli, “Contrast in complex images,” J. Opt. Soc. Am. A 7, 2032–2040 (1990).
    [CrossRef] [PubMed]
  15. Note that the high-frequency half-maximum point of the 1.5-octave Gaussian filter centered at the 4 cycles/face that were used by Schuchard and Rubin is at 6 cycles/face.11,18
  16. J. A. Swets, R. M. Pickett, Evaluation of Diagnostic Systems: Methods from Signal Detection Theory(Academic, New York, 1982), pp. 15–45.
  17. C. E. Metz, P. Wang, H. B. Kronman, “A new approach for testing the significance of differences between ROC curves measured from correlated data,” in Proceedings of the Eighth Conference on Information Processing in Medical Imaging, F. Deconinck, ed. (Nijhoff, The Hague, 1983), pp. 431–445.
  18. R. A. Schuchard, G. S. Rubin, “Effect of band width on discrimination and recognition of band pass filtered faces,” in Annual Meeting, Vol. 18 of OSA Proceedings (Optical Society of America, Washington, D.C., 1989), p. 161.
  19. V. Bruce, “Recognizing familiar faces,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, F. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 107–117.
    [CrossRef]
  20. T. B. Lawton, “Image enhancement filters significantly improve reading performance for low vision observers,” Ophthalmol. Physiol. Opt. 12, 193–200 (1992).
    [CrossRef]
  21. T. B. Lawton, “Improved reading performance using individualized compensation filters for observers with losses in central vision,” Ophthalmology 96, 115–126 (1989).
    [PubMed]
  22. F. L. Engel, Institute for Perception Research, Eindhoven University of Technology, The Netherlands (personal communication, 1993).

1992 (3)

E. Peli, “Limitations of image enhancement for the visually impaired,” J. Opt. Vis. Sci. 69, 15–24 (1992).
[CrossRef]

E. Peli, “Display nonlinearity in digital image processing for visual communication,” Opt. Eng. 31, 2374–2382 (1992).
[CrossRef]

T. B. Lawton, “Image enhancement filters significantly improve reading performance for low vision observers,” Ophthalmol. Physiol. Opt. 12, 193–200 (1992).
[CrossRef]

1991 (1)

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

1990 (1)

1989 (2)

T. B. Lawton, “Improved reading performance using individualized compensation filters for observers with losses in central vision,” Ophthalmology 96, 115–126 (1989).
[PubMed]

R. A. Schuchard, G. S. Rubin, “Face identification of band pass filtered faces by low vision observers,” Invest. Ophthalmol. Vis. Sci. Suppl. 30, 396 (1989).

1986 (1)

T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
[CrossRef]

1985 (2)

L. O. Harvey, G. P. Sinclair, “On the quality of visual imagery,” Invest, Ophthalmol. Vis. Sci. Suppl. 26, 281 (1985).

M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).

1984 (2)

G. S. Rubin, K. Siegel, “Recognition of low pass filtered faces and letters,” Invest. Ophthalmol. Vis. Sci. Suppl. 25, 28 (1984).

E. Peli, T. Peli, “Image enhancement for the visually impaired,” Opt. Eng. 23, 47–51 (1984).
[CrossRef]

1983 (1)

A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
[CrossRef] [PubMed]

1980 (1)

A. P. Ginsburg, “Specifying relevant spatial information for image evaluation and display design: an explanation of how we see certain objects,” Proc. Soc. Inf. Disp. 21, 219–227 (1980).

Arend, L.

E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.

Bruce, V.

V. Bruce, “Recognizing familiar faces,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, F. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 107–117.
[CrossRef]

Burr, D. C.

T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
[CrossRef]

Buzney, S. M.

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

Encke, W.

M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).

Engel, F. L.

F. L. Engel, Institute for Perception Research, Eindhoven University of Technology, The Netherlands (personal communication, 1993).

Fiorentini, A.

A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
[CrossRef] [PubMed]

Ginsburg, A. P.

A. P. Ginsburg, “Specifying relevant spatial information for image evaluation and display design: an explanation of how we see certain objects,” Proc. Soc. Inf. Disp. 21, 219–227 (1980).

Goldstein, R.

E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.

Goldstein, R. B.

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

Harvey, L. O.

L. O. Harvey, G. P. Sinclair, “On the quality of visual imagery,” Invest, Ophthalmol. Vis. Sci. Suppl. 26, 281 (1985).

Hayes, T.

T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
[CrossRef]

Hübner, M.

M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).

Kronman, H. B.

C. E. Metz, P. Wang, H. B. Kronman, “A new approach for testing the significance of differences between ROC curves measured from correlated data,” in Proceedings of the Eighth Conference on Information Processing in Medical Imaging, F. Deconinck, ed. (Nijhoff, The Hague, 1983), pp. 431–445.

Lawton, T. B.

T. B. Lawton, “Image enhancement filters significantly improve reading performance for low vision observers,” Ophthalmol. Physiol. Opt. 12, 193–200 (1992).
[CrossRef]

T. B. Lawton, “Improved reading performance using individualized compensation filters for observers with losses in central vision,” Ophthalmology 96, 115–126 (1989).
[PubMed]

Maffei, L.

A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
[CrossRef] [PubMed]

Metz, C. E.

C. E. Metz, P. Wang, H. B. Kronman, “A new approach for testing the significance of differences between ROC curves measured from correlated data,” in Proceedings of the Eighth Conference on Information Processing in Medical Imaging, F. Deconinck, ed. (Nijhoff, The Hague, 1983), pp. 431–445.

Morrone, M. C.

T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
[CrossRef]

Peli, E.

E. Peli, “Limitations of image enhancement for the visually impaired,” J. Opt. Vis. Sci. 69, 15–24 (1992).
[CrossRef]

E. Peli, “Display nonlinearity in digital image processing for visual communication,” Opt. Eng. 31, 2374–2382 (1992).
[CrossRef]

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

E. Peli, “Contrast in complex images,” J. Opt. Soc. Am. A 7, 2032–2040 (1990).
[CrossRef] [PubMed]

E. Peli, T. Peli, “Image enhancement for the visually impaired,” Opt. Eng. 23, 47–51 (1984).
[CrossRef]

E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.

Peli, T.

E. Peli, T. Peli, “Image enhancement for the visually impaired,” Opt. Eng. 23, 47–51 (1984).
[CrossRef]

Pickett, R. M.

J. A. Swets, R. M. Pickett, Evaluation of Diagnostic Systems: Methods from Signal Detection Theory(Academic, New York, 1982), pp. 15–45.

Rentschler, I.

M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).

Rubin, G. S.

R. A. Schuchard, G. S. Rubin, “Face identification of band pass filtered faces by low vision observers,” Invest. Ophthalmol. Vis. Sci. Suppl. 30, 396 (1989).

G. S. Rubin, K. Siegel, “Recognition of low pass filtered faces and letters,” Invest. Ophthalmol. Vis. Sci. Suppl. 25, 28 (1984).

R. A. Schuchard, G. S. Rubin, “Effect of band width on discrimination and recognition of band pass filtered faces,” in Annual Meeting, Vol. 18 of OSA Proceedings (Optical Society of America, Washington, D.C., 1989), p. 161.

Sandini, G.

A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
[CrossRef] [PubMed]

Schuchard, R. A.

R. A. Schuchard, G. S. Rubin, “Face identification of band pass filtered faces by low vision observers,” Invest. Ophthalmol. Vis. Sci. Suppl. 30, 396 (1989).

R. A. Schuchard, G. S. Rubin, “Effect of band width on discrimination and recognition of band pass filtered faces,” in Annual Meeting, Vol. 18 of OSA Proceedings (Optical Society of America, Washington, D.C., 1989), p. 161.

Sergent, J.

J. Sergent, “Microgenesis of face perception,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, R. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 17–33.
[CrossRef]

Siegel, K.

G. S. Rubin, K. Siegel, “Recognition of low pass filtered faces and letters,” Invest. Ophthalmol. Vis. Sci. Suppl. 25, 28 (1984).

Sinclair, G. P.

L. O. Harvey, G. P. Sinclair, “On the quality of visual imagery,” Invest, Ophthalmol. Vis. Sci. Suppl. 26, 281 (1985).

Swets, J. A.

J. A. Swets, R. M. Pickett, Evaluation of Diagnostic Systems: Methods from Signal Detection Theory(Academic, New York, 1982), pp. 15–45.

Trempe, C.

E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.

Trempe, C. L.

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

Wang, P.

C. E. Metz, P. Wang, H. B. Kronman, “A new approach for testing the significance of differences between ROC curves measured from correlated data,” in Proceedings of the Eighth Conference on Information Processing in Medical Imaging, F. Deconinck, ed. (Nijhoff, The Hague, 1983), pp. 431–445.

Young, G. M.

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

Human Neurobiol. (1)

M. Hübner, I. Rentschler, W. Encke, “Hidden face recognition: comparing foveal and extrafoveal performance,” Human Neurobiol. 4, 1–7 (1985).

Invest, Ophthalmol. Vis. Sci. Suppl. (1)

L. O. Harvey, G. P. Sinclair, “On the quality of visual imagery,” Invest, Ophthalmol. Vis. Sci. Suppl. 26, 281 (1985).

Invest. Ophthalmol. Vis. Sci. (1)

E. Peli, R. B. Goldstein, G. M. Young, C. L. Trempe, S. M. Buzney, “Image enhancement for the visually impaired: simulations and experimental results,” Invest. Ophthalmol. Vis. Sci. 32, 2337–2350 (1991).
[PubMed]

Invest. Ophthalmol. Vis. Sci. Suppl. (2)

G. S. Rubin, K. Siegel, “Recognition of low pass filtered faces and letters,” Invest. Ophthalmol. Vis. Sci. Suppl. 25, 28 (1984).

R. A. Schuchard, G. S. Rubin, “Face identification of band pass filtered faces by low vision observers,” Invest. Ophthalmol. Vis. Sci. Suppl. 30, 396 (1989).

J. Opt. Soc. Am. A (1)

J. Opt. Vis. Sci. (1)

E. Peli, “Limitations of image enhancement for the visually impaired,” J. Opt. Vis. Sci. 69, 15–24 (1992).
[CrossRef]

Ophthalmol. Physiol. Opt. (1)

T. B. Lawton, “Image enhancement filters significantly improve reading performance for low vision observers,” Ophthalmol. Physiol. Opt. 12, 193–200 (1992).
[CrossRef]

Ophthalmology (1)

T. B. Lawton, “Improved reading performance using individualized compensation filters for observers with losses in central vision,” Ophthalmology 96, 115–126 (1989).
[PubMed]

Opt. Eng. (2)

E. Peli, T. Peli, “Image enhancement for the visually impaired,” Opt. Eng. 23, 47–51 (1984).
[CrossRef]

E. Peli, “Display nonlinearity in digital image processing for visual communication,” Opt. Eng. 31, 2374–2382 (1992).
[CrossRef]

Perception (2)

T. Hayes, M. C. Morrone, D. C. Burr, “Recognition of positive and negative band pass-filtered images,” Perception 15, 595–602 (1986).
[CrossRef]

A. Fiorentini, L. Maffei, G. Sandini, “The role of high spatial frequencies in face perception,” Perception 12, 195–201 (1983).
[CrossRef] [PubMed]

Proc. Soc. Inf. Disp. (1)

A. P. Ginsburg, “Specifying relevant spatial information for image evaluation and display design: an explanation of how we see certain objects,” Proc. Soc. Inf. Disp. 21, 219–227 (1980).

Other (8)

J. Sergent, “Microgenesis of face perception,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, R. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 17–33.
[CrossRef]

E. Peli, R. Goldstein, C. Trempe, L. Arend, “Image enhancement improves face recognition,” in Noninvasive Assessment of the Visual System, Vol. 7 of 1989 OSA Technical Digest Series (Optical Society of America, Washington, D.C., 1989), pp. 64–67.

Note that the high-frequency half-maximum point of the 1.5-octave Gaussian filter centered at the 4 cycles/face that were used by Schuchard and Rubin is at 6 cycles/face.11,18

J. A. Swets, R. M. Pickett, Evaluation of Diagnostic Systems: Methods from Signal Detection Theory(Academic, New York, 1982), pp. 15–45.

C. E. Metz, P. Wang, H. B. Kronman, “A new approach for testing the significance of differences between ROC curves measured from correlated data,” in Proceedings of the Eighth Conference on Information Processing in Medical Imaging, F. Deconinck, ed. (Nijhoff, The Hague, 1983), pp. 431–445.

R. A. Schuchard, G. S. Rubin, “Effect of band width on discrimination and recognition of band pass filtered faces,” in Annual Meeting, Vol. 18 of OSA Proceedings (Optical Society of America, Washington, D.C., 1989), p. 161.

V. Bruce, “Recognizing familiar faces,” in Aspects of Face Processing, H. D. Ellis, M. A. Jeeves, F. Newcombe, A. Young, eds. (Nijhoff, Boston, Mass., 1986), pp. 107–117.
[CrossRef]

F. L. Engel, Institute for Perception Research, Eindhoven University of Technology, The Netherlands (personal communication, 1993).

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 (11)

Fig. 1
Fig. 1

Five filters used in our study compared with Schuchard and Rubin’s 4-cycles/face (c/face) filter.

Fig. 2
Fig. 2

Examples of the images used: (a) image containing 100% of 8 cycles/face, (b) image containing 75% of 8 cycles/face (see Fig. 1), (c) image containing 50% of 8 cycles/face, (d) image containing 25% of 8 cycles/face, (e) image containing 0% of 8 cycles/face or 100% of 4 cycles/face.

Fig. 3
Fig. 3

Face-recognition performance as a function of the percentage of the 8-cycles/face content. The ratio of the area under the ROC for the filtered images to the area under the ROC for the original unfiltered images for each subject is represented by one data point. Filled circles, nonsignificant difference between degraded and original images; open circles, significant (p < 0.05) difference between degraded and original images.

Fig. 4
Fig. 4

Four band-enhanced filters used in experiment 2 (see legend) compared with the five low-pass filters used in experiment 1 (thin solid curves); c/face, cycles/face.

Fig. 5
Fig. 5

Averaged contrast spectra (radially averaged amplitude spectrum normalized by mean luminance) from five faces. For the filter enhancing the 8-cycles/face band by a factor of 5 (solid curve), the spectrum of the designed filtered image (long-dashed curve) is compared with the actual spectrum resulting from saturation (short-dashed curve) and with the measured spectrum of the corresponding images enhanced by the nonlinear adaptive enhancement (dotted–dashed curve).

Fig. 6
Fig. 6

Degradation in face-recognition performance as a function of the percentage of the content at (a) 8 cycles/face and (b) 16 cycles/face. The notations follow the same convention as in the caption for Fig. 3. Data from experiment 1 (circles) are compared with the data of experiment 2: squares indicate 1-octave bandwidth, triangles indicate 2-octave bandwidth, and diamonds indicate low-contrast images. The solid curves connect the mean values for the 1-octave conditions.

Fig. 7
Fig. 7

Virtual grids of processed images presented for selection: (a) 1-octave, (b) 2-octave. In each square the top numbers represent the center spatial frequency (in cycles/face) of the bands that were enhanced and the lower numbers (bold) the amplification applied to that band.

Fig. 8
Fig. 8

Examples of the face images enhanced with a 1-octave-wide filter. The filter-center frequencies and the amplification factors used are illustrated in Fig. 7(a). The central four images are the same as those used in experiment 2.

Fig. 9
Fig. 9

Examples of the face images enhanced with a 2-octave-wide filter. The filter-center frequencies and the amplification factors used are illustrated in Fig. 7(b).

Fig. 10
Fig. 10

Results of celebrity-recognition experiments (second part of experiment 3). The improvement ratio with the individually tuned enhancement as a function of face-recognition performance is shown for the two comparison conditions: (a) individual enhancement with the original unenhanced image, (b) individual enhancement with the adaptive enhancement.

Fig. 11
Fig. 11

Radially averaged contrast spectrum of unenhanced face images compared with the spectra of the images enhanced with the adaptive enhancement and with a 2-octave-wide filter centered at 16 cycles/face and amplified by a factor of 5: c/face, cycles/face.

Metrics