Pengenalan tanda tangan menggunakan algoritme VFI5 dengan citra pelatihan tunggal

Abstract

Biometrics is the science of establishing the identity of an individual based on the physical, chemical or behavioral attributes of the person. Today, one of the most widely applied biometric object is hand-written signature. In this research, we tried to identify scanned offline hand-written signature using VFI5 algorithm. The classification method is only using a single image as a training image. The feature used as the input for VFI5 algorithm is the gray level intensity of the image and the signature image dimension used in this research is 60 × 40 pixels which means originally there are 2400 features to compute. Then, we reduce the dimension using imresize function with nearestneighbor interpolation method in Matlab. The reduced image dimensions are 45 x 30, 30 x 20, 15 x 10, and 7 x 5 pixels. The reduced images then classified using VFI5 algorithm, with 84.33%, 78.44%, 59.11%, and 32.89% accuracy respectively for the first, second, third, and fourth reduced dimension. The result of this research also shows that the accuracy of the recognition is influenced by the image dimension. It turned out that the smaller the dimension of the image, the lower the accuracy of the recognition. Smaller image dimension also resulted in less training and testing time.