AI-Powered Object Detection

This paper presents an AI-powered Object Detection system designed to provide real-time environmental awareness using computer vision and machine learning techniques. The system captures live video input through a camera and processes each frame to perform object detection, text recognition using OCR, currency identification, and traffic light detection. It includes a direction and urgency analysis module to determine the position and importance of detected objects, enabling effective decision-making. A priority-based alert queue system is implemented to manage and deliver voice alerts using a text-to-speech engine without overlap. The system supports multiple modes such as walk mode, OCR mode, and currency mode, allowing flexible and context-aware operation. Experimental results demonstrate that the system performs efficiently in real-time scenarios with minimal delay, making it a practical and reliable assistive solution, particularly for visually impaired individuals.