• Dr Nagarathna
• Thanushree M
• U Nandini
• Vasanth Kumar N P
• Srujan Krishna

Visual impairment significantly affects independent mobility, environmental awareness, and accessibility to textual information for millions of individuals worldwide. Traditional assistive systems such as white canes and guide dogs provide only limited contextual understanding and fail to deliver intelligent environmental interpretation. Existing AI-powered assistive technologies frequently depend on cloud computing infrastructure, internet connectivity, or expensive wearable hardware, leading to increased latency, privacy concerns, and high implementation costs. This paper presents SMART EYE; an offline Edge AI-based object detection and voice assistance system designed specifically for visually impaired individuals. The proposed framework integrates YOLOv8-based real-time object detection, EasyOCR-based text extraction, distance estimation, positional awareness, and speech synthesis into a unified accessibility platform. The system performs complete offline inference using local CPU resources without cloud dependency, thereby reducing latency and preserving user privacy. The architecture was implemented as a software-centric prototype using Python, OpenCV, YOLOv8, PyTorch, and EasyOCR on a standard webcam-enabled system. Experimental evaluation demonstrated reliable real-time performance with frame rates between 10–20 FPS under standard operating conditions. The proposed framework establishes a scalable and cost-effective foundation for future wearable Edge AI assistive technologies including smart glasses, mobile deployment, and embedded accessibility systems.

Edge AI, YOLOv8, OCR, Computer Vision, Offline AI, Assistive Technology, Deep Learning, Speech Synthesis, Accessibility Systems

"SMART EYE: A REAL-TIME OFFLINE EDGE AI FRAMEWORK FOR OBJECT DETECTION, TEXT RECOGNITION, AND VOICE GUIDANCE IN VISUALLY IMPAIRED ACCESSIBILITY", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 5, page no.g200-g204, May-2026, Available :http://www.jetir.org/papers/JETIR2605625.pdf

"SMART EYE: A REAL-TIME OFFLINE EDGE AI FRAMEWORK FOR OBJECT DETECTION, TEXT RECOGNITION, AND VOICE GUIDANCE IN VISUALLY IMPAIRED ACCESSIBILITY", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 5, page no. ppg200-g204, May-2026, Available at : http://www.jetir.org/papers/JETIR2605625.pdf