ResQ Lens: AI Powered Smart Glasses for Visually Impaired Individuals

This paper illustrates conceptualization and development of AI-Enabled Smart Glasses that utilize the edge computing capabilities of smartphones to enable visually impaired individuals to read their environment in real-time. The system combines small hardware elements such as a camera, microphone, and bone conduction speakers with a standard eyeglasses frame, while making a wireless connection to a smartphone using Wi-Fi Direct or the cellular network. At the heart of the system is a high-performance, optimized edge AI pipeline running completely on the mobile device without the need for persistent internet connectivity. Real-time processing of the video stream from the camera is realized using optimized ONNX-formatted models for each task, such as object detection (YOLOv5n), optical character recognition (OCR), and captioning the scene. Optional voice command support is also integrated and a Text-to-Speech (TTS) engine is used to provide auditory feedback using bone conduction speakers, thus allowing the user to remain cognizant of their surroundings. The project focuses on portability, low power consumption, and offline operation, which are the best features for wearable AI applications. By using modern smartphone processing the architecture overcomes usage of Raspberry Pi or Arduino, reducing hardware complexity and cost.