In the template matching, correlation coefficient has been frequently used to measure the similarity of gray-scale images. It is one of an important matching procedure to select a template of good quality which has influence on matching result. A temp...
In the template matching, correlation coefficient has been frequently used to measure the similarity of gray-scale images. It is one of an important matching procedure to select a template of good quality which has influence on matching result. A template of good quality, which is generally distinguished by the naked eye, is not always successful in template matching experiment. To solve this problem, we introduce a SNR(signal-to-noise ratio) of template to compare templates each other and relates a SNR of template with correlation coefficient. From this compare templates each other and relates a SNR of tem plate with correlation coefficient. From this relation, we explain how to determine a threshold correlation coefficient that is to know whether a object exists or not and introduce method to measure a SNR of template. Subsequently, we show the fact that confidence of template matching method considering a SNR of template is improved.