Object Detection with Voice Assistant App

Visually impaired individuals often face difficulties in identifying objects and reading text from their surroundings. This paper presents a real-time Android application designed to assist such users through object detection and multilingual text reading. The system integrates the YOLOv8 model (converted into TensorFlow Lite) for efficient on-device object detection and Google ML Kit for Optical Character Recognition (OCR). The extracted results are delivered to the user through a multilingual Text-to-Speech (TTS) engine that supports English, Kannada, and Hindi. A gesture-based interface enables hands-free interaction, where a right swipe activates object detection and a left swipe triggers text reading. The application is developed using Java in Android Studio and functions completely offline, ensuring privacy and accessibility. Experimental evaluation demonstrates that the proposed system achieves accurate detection and fast speech response, providing an effective and affordable AI-based solution for visually impaired individuals.