Identification of Nominal Value and Authenticity of Rupiah Using Support Vector Machine

This study was conducted to create a software that can recognize the value nominal rupiah and its authenticity through image processing by using Support Vector Machine (SVM). The pattern of each rupiah has unique characteristics that distinguish each other, either as a number, the number of zeros and its invisible ink. The difference pattern of the rupiah pieces of money that is the characteristic which will be taken as the data for determining the nominal value and authenticity of the rupiah. Image processing money obtained from two sources, namely the image appears (visible image) derived from 300 dpi scanner, and the image of the invisible images that uses ultraviolet light (UV). The identification process is carried out in this study has a two-stage identification. The first stage is the identification of the nominal value of the rupiah and the second stage is the identification of genuine rupiah. Identification is done by taking the nominal value of a particular pattern of on the scanning results rupiah. The pattern here is the form taken by the nominal rate of the respective pieces of money of dollars. Authenticity identification is done by taking pattern of formed from the UV. The total amount of data is as much as 95 sets, of which 60 of them are used in the training process and the remaining 35 sets are used in the testing process. The simulation results showed the software is able to recognize the nominal value and authenticity of the currency by the level of accuracy of 100% (using gaussian) and 99% (using polynomial). Based on these results it can be concluded that the identification of the nominal value and authenticity of rupiah using SVM with Gaussian kernel can be used appropriately and accurately.