GET THE APP

An Automated Ethiopian Banknote Recognition Model for Aiding Visual Impaired People | SciTechnol

Journal of Computer Engineering & Information Technology.ISSN : 2324-9307

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Short Communication, Jceit Vol: 9 Issue: 6

An Automated Ethiopian Banknote Recognition Model for Aiding Visual Impaired People

Asfaw Alene Shefraw*

Department of Fashion technology, Bahir Dar University, Bahir Dar, Ethiopia

*Corresponding Author: Asfaw Alene Shefraw
Department of Ethiopian institute of textile and Fashion technology, Bahir Dar University, Bahir Dar, Ethiopia
Tel: 251918247040
E-mail: asa4ever2002@mail.com

Received: December 4, 2020 Accepted: December 18, 2020 Published: December 25, 2020

Citation: Shefraw A (2020) An Automated Ethiopian Banknote Recognition Model for Aiding Visual Impaired People J Comput Eng Inf Technol 9:6 .DOI: 10.37532/jceit.2020.9(6).243

Abstract

Even though there is an increasing demand on the use of Master cards and other electronic payment, money continues to be broadly utilized for ordinary exchanges due to its convenience. Nevertheless, the visually impaired people may suffer from recognizing each paper money. This necessitates the improved authenticity verification system that may assist visually disabled or blind people to identify denomination and realize if a paper money is authentic. This paper presents the development of a novel camera-based system that utilizes image processing technique with the aim of assisting visually disabled people to automatically recognize banknotes. The exclusive feature for each denomination specific ROI was extracted, and the recognition models were developed and tested. The proposed method first computes the dominant color of the banknotes. Then, the denomination specific ROI was identified automatically. In the end, the ColourMomentum, dominant color, and GLCM feature were calculated from each ROI. Lastly, the genetic optimization algorithm was applied for reducing the dimension of the feature vector. The proposed denomination specific system is effective in collecting class-specific information and has reliable robustness in managing view point changes like partial occlusion, rotation, zooming, and translation.

Keywords: Ethiopian banknotes; Currency recognition; visually disabled; Denomination specific ROI.

According to World Blind Union (WBU), there were 61 million visually impaired people around the world in 2002, which is about 2.6% of the total population. Among these people, 124 million were registered to have low vision and 37 million were blind. As a matter of fact, visually impaired people face a number of challenges when interacting with the environments

Due to this, much information is visually encoded in their daily life. One specific difficulty that a visually disabled person will face is the tasks of identifying the value of the notes he or she is holding. Currently, printed denominations of Ethiopian currencies’ are 1-Birr, 5-Birr, 10-Birr, 50-Birr, and 100-Birr. With a few recent exceptions, all of the banknotes are different in size, shape that is printed through intaglio printing and color which is inaccessible to people who are blind or significantly visually impaired.

The national bank of Ethiopia which is responsible to print hard currencies in the country uses both intrinsic as well as extrinsic features to design the banknotes. The extrinsic properties accommodate size, width, color, etc. whereas the intrinsic properties include security thread, I.D. mark, number panel, etc. Extrinsic properties do not seem to be enough to acknowledge whether the note is original or fake. Also, currency may get damaged during transportation or exchange. The change within the nature of a picture may be well understood and improved with the assistance of image processing techniques. It enhances a number of the protection features embedded within the image for human interpretation.

One local study that the researcher considers of particular relevance to the current study (because of the similarity within the objectives and system setup) was published by Jegnaw and Yare gal . Within the study, they proposed Ethiopian banknote recognition system using dominant color, the correlation of the hue, saturation, and intensity of the HSV color, SURF, and widen strip ROI. Within the study, a template matching method was considered to spot the patterns of every feature. In step with the experimental analysis, 90.42% recognition rates for genuine Ethiopian currency with a median time interval of 1.68 seconds per banknote were achieved. Nevertheless, the study achieves better result, using the colour feature for originality check isn't valid and robust. Moreover, the template matching classifiers aren't invariant to intensity value change.

Another research project which is particularly relevant to this present study. In the study, they proposed a method based on component-based SURF for the recognition of U.S. dollar banknotes. The study aimed to design a model to be used by visually impaired persons in uncontrolled environments. Although their proposed methods are robust to several conditions, including scale variations, rotation, occlusions, wrinkling of the banknote, etc, the proposed models are not applicable within the case of the presence of severe motion blur. Moreover, their method is much more computationally expensive than the one presented in this paper. From the experimental analysis, they achieved a real recognition result of 100% when the acquired banknote image is of sufficient quality. The artificial intelligence improvements and development of technologies have enabled great advances in the use of artificial vision for the recognition of the value of several currencies, such as Euro banknotes Ethiopian banknotes, U.S. Dollars, Rupees.

Track Your Manuscript