Detecting symmetry in grey level images: The global optimization approach

Yossi Gofman, Nahum Kiryati*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

22 Scopus citations


A method for efficient detection of the dominant local reflectional symmetry in grey level images is described. The general approach is to define a local measure of reflectional symmetry that transforms the symmetry detection problem to an optimization problem, and obtain the symmetric regions by an efficient global optimization algorithm. The symmetry of a 1D function can be measured in the frequency domain as the fraction of its energy that resides in symmetric Fourier basis functions. This approach is extended to two dimensions. Locality can be formally treated in terms of the Gabor decomposition and implemented via soft windowing. The resulting measure is a complicated multimodal function of the location of the center of the supporting region, its size, and the orientation of the symmetry axis. A new probabilistic generic algorithm is applied to the determination of the global maximum of the reflectional symmetry function. Less than one thousand evaluations of the local symmetry measure are typically needed in order to locate the dominant symmetry in natural, wildlife test images.

Original languageEnglish
Title of host publicationTrack A
Subtitle of host publicationComputer Vision
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages6
ISBN (Print)081867282X, 9780818672828
StatePublished - 1996
Externally publishedYes
Event13th International Conference on Pattern Recognition, ICPR 1996 - Vienna, Austria
Duration: 25 Aug 199629 Aug 1996

Publication series

NameProceedings - International Conference on Pattern Recognition
ISSN (Print)1051-4651


Conference13th International Conference on Pattern Recognition, ICPR 1996


Dive into the research topics of 'Detecting symmetry in grey level images: The global optimization approach'. Together they form a unique fingerprint.

Cite this