Symmetry detection by categorization of spatial phase, a model

Publisher Website
Google Scholar
Abstract
Bilateral symmetry of patterns or ornaments is attractive to animals as diverse as birds choosing mates and insects searching for food. Although symmetry detection seems to be widespread the mechanisms remain obscure. It has been suggested that this is a more complex task than establishing the co-linearity of a row of dots, requiring matching of points in the two havles of the image, and is perhaps a corollary of the ability to recognize objects regardless of orientation or position. Here I show that axes of bilateral symmetry can be found by one class of local feature detector. For a mechanism that identifies edges and lines by classifying relative phases in spatial harmonics an axis resembles a line with no contrast. Because asymmetries modify spatial phases they will degrade the signal making the mechanism inherently sensitive to fluctuating asymmetry.

This publication has 22 references indexed in Scilit: