VISOR: An AI-Powered Guiding Shield for Vision

Visually impaired people are handicapped in perceiving and acting within their environment based on limited contextual information. Conventional aid tools like white canes and GPS devices give incomplete solutions but cannot offer real-time descriptive directions. With advances in multimodal artificial intelligence (Al), especially vision-language models, it is now feasible to develop smart, chat-like systems that combine vision, language, and speech. This paper introduces VISOR - Guiding Shield for Vision, a multimodal AI-driven assistive system that processes visual input with computer vision, narrates scenes based on vision-language models, and provides natural audio output through speech synthesis. With its integration of object detection, visual question answering, and speech interfaces, VIBE provides context-aware, real-time assistance to visually impaired users. The architecture is lightweight, portable, and modular, which enables the cost-effective solution with scalability for the future.

• Publication year

  2025

• Venue

  2025 Modern Electronics Devices and Intelligent Communication Systems (MEDCOM)

• Publication date

  2025-12-11

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/medcom67532.2025.11405266
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models

  2023cited by this paper
• PaLM-E: An Embodied Multimodal Language Model

  2023cited by this paper
• Learning Transferable Visual Models From Natural Language Supervision

  2021cited by this paper
• YOLOv4: Optimal Speed and Accuracy of Object Detection

  2020cited by this paper
• Deep Learning

  2016cited by this paper
• Very Deep Convolutional Networks for Large-Scale Image Recognition

  2014cited by this paper
• Neural Machine Translation by Jointly Learning to Align and Translate

  2014cited by this paper
• ImageNet Large Scale Visual Recognition Challenge

  2014cited by this paper

• No citing papers are available for this paper.