<![CDATA[Signatures are personal identities that can be used as evidence of document legality. The use of signs is fairly easy and simple, no expensive electronic equipmentneeded. This is what makes this signature is still common nowdays. The higher needs for transactions in society will give effects to the irresponsible parties. The lack ofsecurity tools for transactions or transactions involving this signature makes the roles of signature is at risk. The Signatures belongs to the authentic holder is supposed to be an important discussion for the sake of society.This research will discuss about signature identification to recognize the authentic holder of the signature. This process consists of two main parts: training and testing.The image size is 256x256 pixels. At the training stage, hand drawings used are the threshold, wavelet Symlet and Coiflet wavelets, normalization and then will begenerated using Backpropagation Neural Network (ANN) algorithm. The testing stage has the same process but at the end of the process will be comparison between theimage data that has been stored with the comparison image. ANN can work optimally when it is trained using the calculated data input, parameters, and number of nodes on the network. It is expected that these wavelets can determine the optimal wavelets and parameters in identification. The range of applied wavelets are Symlet and Coiflet.]]>
