Open AccessFast feature identification for holographic tracking: the orientation alignment transform
Author(s) -
Bhaskar Jyoti Krishnatreya,
David G. Grier
Publication year - 2014
Publication title -
optics express
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.394
H-Index - 271
ISSN - 1094-4087
DOI - 10.1364/oe.22.012773
Subject(s) - convolution (computer science) , holography , optics , feature (linguistics) , orientation (vector space) , artificial intelligence , computer science , hough transform , identification (biology) , computer vision , tracking (education) , pattern recognition (psychology) , physics , mathematics , image (mathematics) , geometry , artificial neural network , psychology , pedagogy , linguistics , philosophy , botany , biology
The concentric fringe patterns created by features in holograms may be associated with a complex-valued orientational order field. Convolution with an orientational alignment operator then identifies centers of symmetry that correspond to the two-dimensional positions of the features. Feature identification through orientational alignment is reminiscent of voting algorithms such as Hough transforms, but may be implemented with fast convolution methods, and so can be orders of magnitude faster.