Smartphone recognition of the U.S. banknotes' denomination, for visually impaired people

Abstract

We have developed the prototype of a system that recognizes the denomination of the largest U.S.A. currency notes in circulation in Ecuador, aimed at visually impaired people. It is capable of reproducing audio messages that announce the denomination of a banknote in front of a smartphone camera by processing each frame of its continuous filming. This work takes its theoretical basis from the Digital Image Processing (DIP) techniques and primarily from the image recognition method known as Eigenfaces, which is based on the Principal Component Analysis (PCA) mathematical theory. Tests on two different Nokia smartphones show an accuracy rate of 99.8% and a processing speed of at least 7 frames per second.

Currency security and forensics: a survey

“…The RGB channel to grayscale is converted [21,22]. These methods suffer from the fundamental problem that colours with similar luminance, differing in hue, such as white and yellow cannot be differentiated when converting to grayscale.…”

“…Thresholding allows applications to identify and extract a target from a background of a distribution of grayscale levels or texture in image objects. The correct threshold value is determined using adaptive thresholding [21]. This method shows robustness to strong illumination changes, this means that the end system is able to be used under different kinds of environments.…”

“…In [21], digital image processing approach concepts like eigen-face and Principle Component Analysis (PCA) is utilised to identify the U.S. banknote denomination for visually impaired people using smartphones.…”

An empirical approach for currency identification

“…RELATED WORK Supporting literature exists focusing on banknote recognition software intended for a variety of applications such as assisting the visually impaired [2], banknote sorting and Automatic Teller Machine (ATM) software [3], and banknote fatigue detection [4]. A trend observed is that an image is acquired, it is pre-processed then classified and finally the result is output.…”

An Iranian Cash Recognition Assistance System For Visually Impaireds

“…In [15] Grijalva et.al proposed a method that has been tested on different Nokia smartphone. The reported accuracy rate is about 99.8% and the speed of processing reported as 7 frames per second.…”