Assistive App for Visually Impaired People
Date
2023
Student ID
Research Supervisor
Editor
Journal Title
Volume
Issue
Journal Title
Journal ISSN
Volume Title
Abstract
Monetary transactions are an indispensable part of our day to day activities. In case of a large amount of cash transaction, human error is a matter of concern. Thus the need for an efficient automated system for currency recognition has become significant these days. The challenge of currency recognition and object detection remains a significant obstacle for individuals with visual impairments. This issue is particularly pronounced in developing nations, where robust currency recognition systems are scarce. Recent research efforts have sought to address this issue, focusing on the complexities posed by the gradual degradation of currency notes over time. Recognizing currency notes has become increasingly complex due to wear and tear. Notably, the development of currency recognition systems adjust to the needs of visually impaired individuals in Asian countries has been relatively limited. To address this challenge, research has been conducted, leading to the forthcoming implementation of a practical application featuring two core components: an Image Classification Module and a Text-toSpeech Module. The primary hurdle in both modules is to enhance accuracy by using deep learning techniques. This initiative represents a crucial step in improving accessibility for visually impaired individuals, especially in regions with limited resources. The trained deep learning model achieved a remarkable 91% accuracy, indicating the convergence of the model during the validation process. This outcome signifies a significant advancement in providing accurate and reliable currency recognition and object detection for visually impaired users
Description
Keywords
Citation
Department Name
Electrical and Computer Engineering
Publisher
North South University