Penggunaan Data Mining dalam Mengklasifikasi Nominal Uang Rupiah dengan Metode Convolutional Neural Network (CNN)

Abstract

This study aims to implement the Convolutional Neural Network (CNN) method to identify the authenticity of Indonesian banknotes through image classification into seven categories (classes). The dataset used consists of banknote images captured under various real-world conditions, including differences in printing quality, degrees of wear or deterioration, lighting variations, and multiple shooting angles to obtain diverse image variations. Through this approach, the system is expected to learn distinctive visual patterns and features of each banknote denomination, including texture, color, and embedded security elements found in genuine currency. After the training process, the model is evaluated to measure its accuracy and frames per second (FPS) as performance indicators for real-time recognition. The results of this research are expected to contribute to the development of effective and efficient image processing technology to assist in the automatic classification and detection of Indonesian banknote authenticity, thereby minimizing human error and enhancing security in financial transactions.