High speed paper currency recognition by neural networks

Abstract

In this paper a new technique is proposed to improve the recognition ability and the transaction speed to classify the Japanese and US paper currency. Two types of data sets, time series data and Fourier power spectra, are used in this study. In both cases, they are directly used as inputs to the neural network. Furthermore, we also refer a new evaluation method of recognition ability. Meanwhile, a technique is proposed to reduce the input scale of the neural network without preventing the growth of recognition. This technique uses only a subset of the original data set which is obtained using random masks. The recognition ability of using large data set and a reduced data set are discussed. In addition to that the results of using a reduced data set of the Fourier power spectra and the time series data are compared.

Robust and Effective Component-Based Banknote Recognition for the Blind

“…Kosaka and Omatu proposed the learning vector quantization (LVQ) method to recognize 8 kinds of Italian Liras [16, 17]. Most banknote recognition methods employed neural network techniques for classification [13, 16-18, 31-33]. …”

“…For instance, the average recognition rate of the algorithm [27] based on SIFT is 76%. Some neural network based banknote recognition systems achieved recognition rate no large than 95% [16, 17, 31-33]. Because of specific design, the testing images in these references were captured in highly constrained environments, such as a transaction machine or a specific sensor system which are not portable [5, 31, 33].…”

“…Recent advances in theories, sensors, and embedded computing hold the promise to enable computer vision technique to address their needs. Although a number of studies on camera-based bill recognition have been published in literatures [3, 9, 10, 13, 16-18, 31-33], some of them are restricted to specific and standard environment. For example, the whole bill must be visible without occlusion, wrinkles, etc.…”

Invariant Features Extraction for Banknote Classification

“…Therefore, the coordinates of pixels in each feature block is not sensitive to the rotation. The average feature element that is extracted by (2) and (3) in a feature block is invariant of rotation. …”

Computer vision based currency classification system

“…In above text the methods discussed were mainly focused on the image recognition. Takeda et al [2] have devised approach for verification, for that statistical method is being employed. This classification was giving good responses but it was not able to solve when the data is heavily co-related.…”