Low-Cost Multifunctional Assistive Device for Visually Impaired Individuals
Open Access
- 24 March 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 56326-56337
Abstract
Visually impaired individuals face challenges in mobility, object recognition, and text reading. Existing assistive tools provide partial solutions but lack full integration. This paper presents a low-cost, multifunctional assistive device in the form of wearable glasses, integrating real-time text recognition, obstacle detection, and remote assistance. Built around a Raspberry Pi Zero 2W, the system features an Arducam IMX519 autofocus camera and a tactile button interface for user-friendly control. Real-time audio feedback enhances situational awareness, promoting greater independence. The device was tested at the Rehabilitation Center for the Visually Impaired in Levoča. Performance evaluations confirm high text recognition accuracy and improved navigation support. A comparative analysis highlights its affordability and functionality over commercial alternatives. While effective, the device requires processing speed optimization and hardware miniaturization. Future improvements will explore haptic feedback and AI-driven scene analysis to enhance usability. This research contributes to affordable and inclusive assistive technology, improving accessibility for visually impaired individuals.Keywords
Funding Information
- Slovak Grant Nadácia Tatra Banky
This publication has 30 references indexed in Scilit:
- The global burden of blindnessCurrent Opinion in Ophthalmology, 2024
- A Qualitative and Quantitative Analysis of Research in Mobility Technologies for Visually Impaired PeopleIEEE Access, 2023
- SGBANet: Semantic GAN and Balanced Attention Network for Arbitrarily Oriented Scene Text Recognition2022
- CNN-Based Object Recognition and Tracking System to Assist Visually Impaired PeopleIEEE Access, 2022
- IoT-Based Smart Blind Stick2021
- A pragmatic convolutional bagging ensemble learning for recognition of Farsi handwritten digitsThe Journal of Supercomputing, 2021
- Analysis and comparison of machine learning classifiers and deep neural networks techniques for recognition of Farsi handwritten digitsThe Journal of Supercomputing, 2020
- Coordinate CNNs and LSTMs to categorize scene images with multi-views and multi-levels of abstractionExpert Systems with Applications, 2019
- Accurate Scene Text Recognition Based on Recurrent Neural Network2015
- A Visual SLAM System on Mobile Robot Supporting Localization Services to Visually Impaired People2015