Deploying YOLO Models on HoloLens 2: Balancing Performance and Efficiency in AR Applications

Abstract

This study explores deploying efficient YOLO object detection models on the Microsoft HoloLens 2, an advanced augmented reality (AR) device. The HoloLens 2's limited computational power, memory, and energy efficiency present challenges for implementing AI models like YOLO. We focus on optimizing FPS and model performance to ensure seamless AR interactions, addressing these resource constraints while enhancing the user experience.