Invariant Measures of Image Features From Phase Information
If reliable and general computer vision techniques are to be developed
it is crucial that we find ways of characterizing low-level image
features with invariant quantities. For example, if edge significance
could be measured in a way that was invariant to image illumination
and contrast, higher-level image processing operations could be
conducted with much greater confidence. However, despite their
importance, little attention has been paid to the need for invariant
quantities in low-level vision for tasks such as feature detection or
In my research I have developed a number of invariant low-level
image measures for feature detection, local symmetry/asymmetry
detection, and for signal matching. These invariant quantities are
developed from representations of the image in the frequency
domain . In particular, phase data is used as the fundamental
building block for constructing these measures. Phase congruency can
be used as an illumination and contrast invariant measure of feature
significance. This allows edges, lines and other features to be
detected reliably, and fixed thresholds can be applied over wide
classes of images. Points of local symmetry and asymmetry in images
can be detected from the special arrangements of phase that arise at
these points, and the level of symmetry/asymmetry can be characterized
by invariant measures. It is also possible to perform signal matching
using correlation of local phase and amplitude information. This
approach allows reliable phase based disparity measurements to be
made, overcoming many of the difficulties associated with scale-space
Kovesi, P. D. ``A Dimensionless Measure of Edge Significance'', The
Australian Pattern Recognition Society, Conference on Digital Image
Computing: Techniques and Applications, Melbourne, 4-6 December
1991. pp. 281-288.
Kovesi, P. D. ``A Dimensionless Measure of Edge Significance from
Phase Congruency Calculated via Wavelets'', The First New Zealand
Conference on Image and Vision Computing , Auckland, 16-18 August
1993, pp. 87-94.
Kovesi, P. D. and Trevelyan, J. P. ``Using Visual Doppler Effects to
Deduce Image Motion'', The Australian Pattern Recognition
Society, Conference on Digital Image Computing: Techniques and
Applications, Sydney, 8-10 December 1993. pp. 493-500.
- C. J. Pudney, P. D. Kovesi and B. J. Robbins, ``Feature Detection
Using Oriented Local Energy for 3D Confocal Microscope
Images''. Proceedings of The International Computer Science
Conference, ICSC '95, pp 274-282, December 1995. Hong
Kong. Published by Springer Verlag.
- C. J. Pudney, P. D. Kovesi and B. J. Robbins, ``A 3D Local Energy
Surface Detector for Confocal Microscope Images''. Proceedings of
The Third Australian and New Zealand Conference on Intelligent
Information Systems, pp 7-12, November 1995. Perth,
Western Australia. Published by IEEE.
Peter Kovesi. ``Image Correlation From Local Frequency Information''.
Proceedings of The Australian Pattern Recognition Society
Conference: DICTA'95. December 1995. Brisbane. pp 336-341.
- C. J. Pudney, M. J. Robins, B. J. Robbins and P. D. Kovesi, ``Surface
Detection in 3D Confocal Microscope Images via Local Energy and Ridge
Tracing''. The Journal of Computer Assisted Microscopy, vol. 8,
no. 1, 1996. pp 5-20.
- Peter Kovesi, "Symmetry and Asymmetry From Local Phase"
AI'97, Tenth Australian Joint Conference on Artificial Intelligence.
2 - 4 December 1997. Proceedings - Poster Papers. pp 185-190.
- Peter Kovesi, "Image Features From Phase Congruency". Videre:
A Journal of Computer Vision Research. MIT Press. Volume 1, Number
3, Summer 1999
- Peter Kovesi, "Phase Preserving Denoising of Images".
The Australian Pattern Recognition Society Conference:
DICTA'99. December 1999. Perth WA. pp 212-217
- Peter Kovesi, "Phase congruency: A low-level image invariant
". Psychological Research Psychologische
Forschung. Springer-Verlag. Volume 64, Number 2, 2000 pp 136-148
- Chih W. Koh and Peter Kovesi, "Rotating the Impossible
Retangle". Leonardo. MIT Press. Volume 34, Number 3,
2001. p 197.
- Kristin J. McLoughlin, Philip J. Bones and Peter D. Kovesi,
"Detection of microcalcifications in digital mamograms", Proceedings
of Image and Vision Computing '01, University of Otago,
Dunedin, New Zealand, 26-28 November 2001. pp 259-264.
- Peter Kovesi, "Edges Are Not Just Steps". Proceedings of
ACCV2002 The Fifth Asian Conference on Computer Vision,
Melbourne Jan 22-25, 2002. pp 822-827.
- Kristin J. McLoughlin, Philip J. Bones, and Peter
D. Kovesi. "Connective tissue representation for detection of
microcalcifications in digital mammograms," The SPIE
Medical Imaging Conference, San Diego, 23-28 February 2002.
- Peter Kovesi. "Image Features From Phase Congruency". Department
of Computer Science Technical Report 95/4, June 1995.
technical report has been superseded by the paper published in
Videre (a link to this paper is available above).
MATLAB Code For Calculating Phase Congruency:
phasecong.m MATLAB code for calculating
Phase Congruency and Phase Symmetry/Asymmetry.
nonmaxsup.m MATLAB code for non-maxima suppression.
hysthresh.m MATLAB code for hysteresis thresholding.
Example MATLAB session:
>> im = imread('picci.tif'); % Read the image
>> image(im); % Display the image
>> [pc orient ft] = phasecong(im); % Calculate phase congruency.
>> imagesc(pc); % Display the phase congruency image
>> nonmax = nonmaxsup(pc, orient, 1.5); % Perform non-maximal suppression
>> features = hysthresh(nonmax, 0.3, 0.15); % Hysteresis threshold
>> imagesc(features); % Look at the result
Raw phase congruency image obtained from phasecong.m. Phase
congruency is a dimensionless quantity, its value ranges between 0 and
1. Step and line features are marked equally well. Note how each
whisker has been marked as a single feature.
Phase Congruency edge map obtained after non-maxiam suppression
(using nonmaxsup.m) and hysteresis thresholding between phase
congruency values of 0.3 and 0.15 (using hysthresh.m).
For comparison here is the raw Canny edge strength image (sigma =
1). Note how edges are marked on each side of each whisker.
Raw phase congruency image. Notice the response of the detector
in the low contrast regions. Compare this to the output of the Canny
Phase congruency edge map obtained with non-maxima suppression and
hysteresis thresholding between phase congruency values of 0.3 and
Raw Canny edge strength image (sigma = 1). Note how the response of
the Canny detector collapses in the low contrast regions.
Raw phase symmetry image (obtained using phasecong.m). Phase
symmetry is a dimensionless quantity, its value ranges between 0
and 1. It detects symmetry of features independent of contrast.