CLASSIFICATION OF SIGNATURES USING TEMPLATE MATCHING METHOD

Abstract

Template Matching is one of the methods used for digital image processing, usually used to recognize the shape or pattern of an image. The shape or pattern that is often used to be recognized is in the form of character images, letters, numbers, or fingerprints. In the research conducted, signature pattern recognition was made using Template Matching for signature classification. Signature is chosen in research conducted with the aim of knowing whether the signature can be recognized using the Template Matching in addition to character images of letters, numbers, or fingerprints. Template Matching works by matching each pixel in the image matrix that has been digitally processed with the reference image (template) and because Template Matching is an applied method of convolutional technique, Template Matching combines two numbers to produce a third number series, so that the correlation coefficient (r) of the Template Matching will be obtained between -1 and +1. The results of the trials carried out show that the signature pattern recognition with Template Matching can recognize the signature image tested with a recognition accuracy rate of 96% with as many as 100 signature images.