Normalized cross correlation (NCC) has been used extensively for many machine vision applications, but the traditional normalized correlation operation does not meet speed requirements for time-critical applications. In this paper, we propose a fast NCC computation for defect detection. A sum-table scheme is utilized, which allows the calculations of image mean, image variance and cross-correlation between images to be invariant to the size of template window. For an image of size <i>M × N</i> and a template window of size <i>m × n</i>, the computational complexity of the traditional NCC involves 3 ċ <i>m ċ n ċ M ċ N</i> additions/subtractions and 2 ċ <i>m ċ n ċ M ċ N</i> multiplications. The required numbers of computations of the proposed sum-table scheme can be significantly reduced to only 18 ċ <i>M ċ N</i> additions/subtractions and 2 ċ <i>M ċ N</i> multiplications.
Description
Fast normalized cross correlation for defect detection
%0 Journal Article
%1 Tsai:2003:FNC:945085.945096
%A Tsai, Du-Ming
%A Lin, Chien-Ta
%C New York, NY, USA
%D 2003
%I Elsevier Science Inc.
%J Pattern Recogn. Lett.
%K detection feature
%N 15
%P 2625--2631
%R 10.1016/S0167-8655(03)00106-5
%T Fast normalized cross correlation for defect detection
%U http://dx.doi.org/10.1016/S0167-8655(03)00106-5
%V 24
%X Normalized cross correlation (NCC) has been used extensively for many machine vision applications, but the traditional normalized correlation operation does not meet speed requirements for time-critical applications. In this paper, we propose a fast NCC computation for defect detection. A sum-table scheme is utilized, which allows the calculations of image mean, image variance and cross-correlation between images to be invariant to the size of template window. For an image of size <i>M × N</i> and a template window of size <i>m × n</i>, the computational complexity of the traditional NCC involves 3 ċ <i>m ċ n ċ M ċ N</i> additions/subtractions and 2 ċ <i>m ċ n ċ M ċ N</i> multiplications. The required numbers of computations of the proposed sum-table scheme can be significantly reduced to only 18 ċ <i>M ċ N</i> additions/subtractions and 2 ċ <i>M ċ N</i> multiplications.
@article{Tsai:2003:FNC:945085.945096,
abstract = {Normalized cross correlation (NCC) has been used extensively for many machine vision applications, but the traditional normalized correlation operation does not meet speed requirements for time-critical applications. In this paper, we propose a fast NCC computation for defect detection. A sum-table scheme is utilized, which allows the calculations of image mean, image variance and cross-correlation between images to be invariant to the size of template window. For an image of size <i>M × N</i> and a template window of size <i>m × n</i>, the computational complexity of the traditional NCC involves 3 ċ <i>m ċ n ċ M ċ N</i> additions/subtractions and 2 ċ <i>m ċ n ċ M ċ N</i> multiplications. The required numbers of computations of the proposed sum-table scheme can be significantly reduced to only 18 ċ <i>M ċ N</i> additions/subtractions and 2 ċ <i>M ċ N</i> multiplications.},
acmid = {945096},
added-at = {2012-10-15T12:39:57.000+0200},
address = {New York, NY, USA},
author = {Tsai, Du-Ming and Lin, Chien-Ta},
biburl = {https://www.bibsonomy.org/bibtex/2652292b1c77c0fdac6511e86c84c32eb/daill},
description = {Fast normalized cross correlation for defect detection},
doi = {10.1016/S0167-8655(03)00106-5},
interhash = {cf4ab0d1bede62eaa1e201b14e9820e9},
intrahash = {652292b1c77c0fdac6511e86c84c32eb},
issn = {0167-8655},
issue_date = {November 2003},
journal = {Pattern Recogn. Lett.},
keywords = {detection feature},
month = nov,
number = 15,
numpages = {7},
pages = {2625--2631},
publisher = {Elsevier Science Inc.},
timestamp = {2012-10-15T12:39:57.000+0200},
title = {Fast normalized cross correlation for defect detection},
url = {http://dx.doi.org/10.1016/S0167-8655(03)00106-5},
volume = 24,
year = 2003
}