Title :
A computing method of double linear correlation for mirror image matching
Author_Institution :
Decision Sci. & Technol, Tokyo Inst. of Technol., Tokyo, Japan
Abstract :
A computing method of double linear correlation is proposed in order to estimate the location in which the part of an image matches another image even if one of the images is mirrored across the vertical, horizontal, or both axes. 1-D double linear correlation (DLC) is extended to 2-D DLC. The reduction of the number of zero-padding is considered. The computational complexity of the proposed method is lower than that of an FFT approach by more than 50% the number of operations on real numbers.
Keywords :
fast Fourier transforms; image matching; FFT approach; computing method; double linear correlation; mirror image matching; zero-padding; Computational complexity; Convolution; Correlation; Discrete cosine transforms; Image matching; Indium tin oxide; Mirrors; Discrete cosine transform; image matching; linear correlation; mirror image;
Conference_Titel :
Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on
Conference_Location :
Florence
DOI :
10.1109/ICASSP.2014.6854114
