Hybrid discriminative models for banknote recognition and anti-counterfeit

Abstract

Nowadays, advanced technology has played an important task in circulation of anti-counterfeit notes economy. It is essential that requires an efficient solution to detect fake banknotes. This paper proposes an approach for recognition of paper currency based fundamental image processing using deep learning for feature extraction and recognition. Deep neural network techniques have dramatically become the state of the art in image processing. The high capacity of traditional techniques on currency image dataset has been impeded because of varieties of the appearance of the banknotes. This paper focuses recognition face value and anti-counterfeit based on banknote appearance. The proposed method can be applied to recognize many kinds of the denomination or face values as well as the national currencies. The contribution studies a new approach based on sequential deep neural network and data augmentation for improving accuracy. First, the deep neural network is constructed using several inceptions with different parallel convolutional operations which support reducing consuming time. Second, image augmentation of training dataset generates a larger data enough for deep neural network learning. This proposed task is aimed to address the small data problem. It is utilized for enhancing the capabilities of deep learning. Experimental results illustrate that the proposed method is applicable to the real application with enhances performance to 99.97% accuracy rate.