Money Recognition for the Visually Impaired: A Case Study on Sri Lankan Banknotes

Akshaan Bandara · 2025

Read Money Recognition for the Visually Impaired: A Case Study on Sri Lankan Banknotes by Akshaan Bandara (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-14267.txt.

Introduction

Akshaan Bandara builds a banknote reader in 2025 to help blind people in Sri Lanka handle cash with confidence. Telling notes apart by touch is unreliable, and errors can cost money and independence. Past aids were often awkward to use, so Bandara aims for a stand-alone system with a simple interface.

A custom image set of Sri Lankan notes was used to fine-tune an object-spotting model. The model scored very high accuracy on held-out images and worked well in real-world trials. Blind users could identify note values quickly and accurately with the device. Bandara presents the system as a step toward safer, more independent daily payments.

Catalog record

CREATORAkshaan Bandara
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-14267

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Money Recognition for the Visually Impaired: A Case Study on Sri Lankan Banknotes Akshaan Bandara Abstract Currency note recognition is a critical accessibility need for blind individuals, as identifying banknotes accurately can impact their independence and security in financial transactions. Several traditional and technological initiatives have been taken to date. Nevertheless, these approaches are less user-friendly and have made it more challenging for blind people to identify banknotes. This research proposes a user-friendly stand-alone system for the identification of Sri Lankan currency notes. A custom-created dataset of images of Sri Lankan currency notes was used to fine-tune an EfficientDet model. The currency note recognition model achieved 0.9847 AP on the validation dataset and performs exceptionally well in real-world scenarios. The high accuracy and the intuitive interface have enabled blind individuals to quickly and accurately identify currency denominations, ultimately encouraging accessibility and independence. I Introduction A disability is a condition that hinders an individual’s ability to perform certain tasks or interact with others. The most common types of disabilities are Vision Impairment, Speech Impairment, and Cognitive Impairment. The World Health Organization (WHO) estimates that at least 1.3 billion of the world’s population suffer from an impairment [ 1 ] . In Sri Lanka, approximately 1.6 million people are disabled [ 2 ] . Vision impairment is the most common disability among known disabilities. According to WHO, at least 295 million individuals suffer from severe visual impairment. Moreover, at least 43 million of them are blind [ 3 ] . It is estimated that 3.9 million people in Sri Lanka suffer vision loss, of which 89,000 are blind [ 4 ] . In Sri Lanka, banknotes are used for the majority of financial transactions. A normal person has no difficulty identifying banknotes. However, it is challenging for those with vision impairments to identify banknotes. As a result, people frequently take advantage of their disability to defraud them. This issue needs to be given more focus. It is, however, a matter that receives minimal attention. A thorough literature review was conducted on current initiatives and research in Sri Lanka and other countries. The physical approaches include embossed dots on the banknotes. But, those dots tend to wear off. Moreover, some research has already been conducted using Deep Learning (DL) and Machine Learning (ML). The solutions proposed in those works can accurately identify the banknotes. However, the majority of the work is not user-friendly. Hence, it is necessary to solve these constraints. The goals of this study are established based on those limitations. Thus, the research aims to design, develop, and evaluate a user-friendly interface that will allow Visually Impaired People (VIP) to accurately and efficiently identify Sri Lankan money (specifically banknotes) in diverse backgrounds. Since Sri Lanka is a developing country, many blind people, especially in the rural areas, might not have access to smartphones. Additionally, those blind people don’t often use physical currency. However, VIP in urban areas use both physical currency and smartphones for their daily activities. So, VIP who utilize cash and smartphones for their everyday activities will be the intended audience for the proposed system. As a result, the most precise solution will be a mobile application. More precisely, an application that can both identify Sri Lankan banknotes in real time and function as a stand-alone system should be developed. Thus, an efficient and accurate Object Detection (OD) model could be utilized on a dataset that has varying lighting and background images to achieve this task. Since there is not much image data for the Sri Lankan banknotes, data augmentation techniques could be used to expand the image dataset. Eventually, it should improve the accuracy of the model. The rest of the paper is as follows: Section II reviews the previous work and outlines the research gap. A detailed description of the methodology is provided in Section III. The results are presented in Section IV, and the paper concludes with an overview of future work in Section V. II Literature review II-A Related work The most common way VIP use banknotes is by placing those in various parts of their wallets. A relative does this task on behalf of the individual. Thus, it is convenient for the blind person to handle banknotes. Nonetheless, the visually impaired individual should always depend on others to put the money in the appropriate places. Furthermore, the Central Bank of Sri Lanka (CBSL) has taken initiatives to aid VIP in identifying denominations. So, as illustrated in Fig. 1 , heavily printed dots appear on the left side of the currency note. Fig. 1: Heavily printed dots on Sri Lankan banknotes [ 5 ] . Although it is an important step towards helping the VIP, the heavily printed dots wear out with use. As a result, the VIP finds it increasingly difficult to use banknotes. Due to recent advancements in Artificial Intelligence (AI), some excellent work has been done to address this issue. Nevertheless, more research needs to be done on Sri Lankan banknotes. Hence, the work that has been done in Sri Lanka and other countries is taken into consideration. Furthermore, only the work done in the last five years was considered for the literature survey of this research. A study on the Sri Lankan notes and coins is presented in [ 6 ] . Convolutional Neural network (CNN) was used to classify the denominations. The note detection accuracy was 0.6667 and the coin detection accuracy was 0.6970. The low accuracy might have occurred due to CNN training from scratch. However, in their approach, an image should be captured and then the inference should be done. Thus, it can lead to a less user-friendly application. Another research on the Sri Lankan banknotes is presented in [ 7 ] . A CNN-based Transfer Learning (TL) model was fine-tuned to classify the Sri Lankan banknotes. Popular pre-trained models such as MobileNetV2, NASNetMobile, EfficientNetB0, DenseNet121, DenseNet201, ResNet50V2, VGG16, and VGG19 were trained and tested to identify the best-performing model. As a result, DenseNet121 performed well by achieving 98% of the test dataset. A focus mechanism is used to recognize the denominations. Hence, the model will detect the currency note when it is approximately 10 cm from the camera. The solution is less user-friendly from the perspective of a blind person, despite the model’s exceptional accuracy. There has not been much research done on this area in recent times in Sri Lanka. Thus, research done in other countries is also considered to derive the research gap. A study was done on the Bangladeshi banknotes utlizing the Oriented FAST and Rotated BRIEF (ORB) algorithm [ 8 ] . OpenCV was used to generate the key points, which were then stored in a database. In production, when an image is received, it is converted to grayscale to eliminate the lighting conditions. The ORB keypoint detection algorithm was then used to identify the key points, and the ORB descriptor extractor was used to extract them. The brute force Hamming algorithm was then utilized to match the key points. The highest number of matching points is finally measured. The average matching time was 0.17s, and the accuracy was 100%. This approach is, however, not a user-friendly solution. Another research was done for

Read the complete text