Real time object detection and interaction in at using deep learning

Abstract—Augmented Reality (AR) has made extensive progress in various applications such as gaming, medical, and industrial automation. Support from deep learning-based real-time object detection offers improved user engagement and environment insight in AR environments. This article introduces a solution using Convolutional Neural Networks (CNNs) and existing state-of-the-art object detectors like YOLO (You Only Look Once) and Faster R-CNN for real-time object identification. The system handles camera input, real-time object detection, and interactive AR overlays based on object properties and user input. Model quantization and hardware acceleration through GPUs and TPUs minimize latency. Experimental tests show the efficiency of the proposed method in accuracy, detection speed, and interactive responsiveness. The findings promise enhanced real-time AR usage, opening the door to smart, context-aware augmented surroundings. Keywords—Real-time object detection, Augmented Reality, Deep Learning, YOLO, Faster R-CNN, Convolution.