Open AccessLocal orientation coherence based segmentation and boundary-aware diffusion for discontinuous fringe patterns
Author(s) -
Haixia Wang,
Qian Kemao
Publication year - 2016
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.24.015609
Subject(s) - discontinuity (linguistics) , coherence (philosophical gambling strategy) , computer science , boundary (topology) , orientation (vector space) , segmentation , artificial intelligence , optics , noise reduction , computer vision , algorithm , pattern recognition (psychology) , geometry , mathematics , physics , mathematical analysis , statistics
Fringe patterns with noise and discontinuity are often encountered but difficult to analyze. Discontinuity-detectable and boundary-aware processing techniques are demanded. A local orientation coherence based fringe segmentation (LOCS) method and its cooperation with boundary-aware coherence enhancing diffusion (BCED) for discontinuous fringe pattern denoising are proposed in this paper. The LOCS method has three steps. First, as orientation coherence indicated by structure tensors is informative to describe fringe structures, it is selected for discontinuity recognition. Due to the complexity of the discontinuity problem, the detected boundary often has missing parts and is not very accurate. Boundary completion by cubic splines and boundary refinement based on partial structure tensors are further performed as the second and third steps, respectively. Subsequently, the BCED method is developed to adapt the original CED to fringe segments with irregular boundaries. Simulated and experimental fringe patterns are tested and successful results have been obtained.