Open AccessMotion‐blurred SIFT invariants based on sampling in image deformation space and univariate search
Author(s) -
Fu Bo,
Guo Hao,
Zhao Xilin,
Chang Yufang,
Li Bo,
He Li
Publication year - 2016
Publication title -
iet computer vision
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.38
H-Index - 37
eISSN - 1751-9640
pISSN - 1751-9632
DOI - 10.1049/iet-cvi.2015.0076
Subject(s) - scale invariant feature transform , artificial intelligence , computer vision , invariant (physics) , matching (statistics) , mathematics , scale space , computer science , feature extraction , pattern recognition (psychology) , image (mathematics) , image processing , statistics , mathematical physics
Scale‐invariant feature transform (SIFT) operator is a widely used algorithm to detect local features in images, yet applying this algorithm for motion‐blurred image matching is difficult and inefficient. To resolve this issue, this study presents a motion‐blurred invariant SIFT algorithm that is based on sample matching in an image deformation reconstructed space. First, the motion‐blurred equation is deduced and its controlling parameters to reconstruct the deformation space are discretised. Second, the authors matched samples in the motion‐blurred space with the blurred image to identify maximal matching points and optimal parameters. In order to improve the searching efficiency, the authors used a univariate search technique combined with a variable step hill‐climbing method to determine the optimal matching. Together, the experimental results show that this improved algorithm has superior matching performance for motion‐blurred images.