Computer Vision Empowered Assistive Technology for Blind People
A cost-effective mobile application is being developed to empower visually impaired individuals by enhancing indoor navigation. This innovative solution leverages advanced computer vision technology, specifically the YOLO v9 object detection model, to provide an affordable alternative to expensive assistive devices. The application delivers real-time assistance, helping users avoid obstacles and navigate safely. Using the smartphone’s camera, the system captures video and processes it through the YOLO v9 model, chosen for its speed and accuracy in real-time object detection. The model is pre-trained to recognize common indoor obstacles such as furniture, walls, doors, and stairs. By identifying these barriers, the app enhances the user’s confidence in navigating complex indoor environments. Feedback is provided through dual channels: audio alerts and haptic feedback. Clear audio cues like "Chair on the left" or "Wall ahead" help users adjust their movements, while haptic feedback provides tactile proximity cues through vibrations that intensify as users approach obstacles. This dual-feedback system ensures usability in noisy environments or when audio guidance is impractical. The app prioritizes accessibility and ease of use, featuring a simple interface for quick activation. The integration of YOLO v9 and real-time feedback creates a practical tool for visually impaired individuals, enabling them to navigate with greater confidence and independence. By offering an affordable and effective solution, this application bridges the gap between advanced assistive technology and inclusivity, improving the quality of life for visually impaired individuals worldwide.
Authors
- P. Mathivanan
- Е. С. Брискин (ORCID: https://orcid.org/0000-0002-6409-4208)
- N Abishek
- Yabez Davidraj P
Institutions
- Karunya University (IN)
Publication Details
- Published
- 2024-12-12
- DOI
- https://doi.org/10.1109/icercs63125.2024.10895124
- Citations
- 2
- Primary Topic
- Tactile and Sensory Interactions
- Type
- article
- Field-Weighted Citation Impact
- 0.69