Logic operation-based template matching algorithm for one-dimensional signals

The normalised form of cross-correlation is a key technique used in template matching problems related to locating and recognising a signal pattern. Since it does not have a simple frequency-domain expression, the computation of the normalised cross-correlation (NCC) is performed in the spatial domain. However, the traditional NCC operation cannot cope with the growing speed requirements of recent time-critical applications. A new NCC-based template matching technique is presented in order to replace expensive multiplications with simpler logic operations. This algorithm is effective in cases where the signal-to-noise-ratio (SNR) is known or in those in which a direct SNR estimation is available. Although the algorithm requires SNR information, it encompasses a wide range of scaling factors and is robust to errors in SNR estimation. Experimental results show that the new template matching method can be used to determine the degree of similarity between test and reference signals with significant improvements over those of traditional methods.