Communication Dans Un Congrès Année : 2022

Computer Vision and Deep Learning-Based support system for Visually Impaired People in an outdoor environment

Résumé

Visual assistance systems are still severely constrained by sensor capabilities, form factor, battery power consumption, computational resources and the use of traditional computer vision algorithms. The challenge is to find innovative solutions that encompass such limitations and solve mobility problems of visual impaired persons (VIP), thus improving their quality of life through using new technologies and AI-Based algorithms. Smartphones and wearable devices, together with an ever-growing uptake of deep learning, computer vision, the Internet of Things (IoT), and virtual and augmented reality are key technologies in this context. In this paper, we propose to design and implement a novel visual assistance system that relies on deep learning and edge processing to perform advanced perception tasks on a Raspberry pi 4, that-is a cost-effective, low power mobile computing platform. The system accomplishes several tasks. It detects objects in the nearby environment of the VIP, and transforms text to sound. It achieves also obstacle avoidance through an ultra-sensor using sound alerts. It can also communicate with family members through a web server. A GPS module helps track the VIP anytime anywhere through this web server. The proposed system design incorporates an AI-based voice interface that allows for user-friendly interaction and control with an acceptable latency. It is also shown that the proposed prototype is a simple low cost, portable and unobtrusive visual assistance device.

Fichier principal
Vignette du fichier
Computer_Vision_and_Deep_Learning_Based_support_system_for_Visually_Impaired_People_in_an_outdoor_environment.pdf (250.24 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04404843 , version 1 (19-01-2024)

Licence

Identifiants

  • HAL Id : hal-04404843 , version 1

Citer

El-Mehdi Baahmed, Ahmed Rafik, Sarah Ouarab, Meriem Smati, Randa Cheima. Computer Vision and Deep Learning-Based support system for Visually Impaired People in an outdoor environment. Colloque sur les Objets et systèmes Connectés 2022, May 2022, Dakar, France. ⟨hal-04404843⟩

Collections

181 Consultations
188 Téléchargements

Partager

  • More