Computer Vision on Edge Devices - Real-Time Object Detection for Retail Applications

Explore how computer vision transforms retail experiences through real-time object detection on resource-constrained edge devices in this conference talk from NDC Oslo 2025. Discover the practical challenges of implementing AI-powered self-checkout systems in high-volume supermarkets, where advanced models must recognize tens of thousands of items while operating on small, low-cost hardware with limited storage and compute capabilities. Learn how NorgesGruppen addresses the balance between creating frictionless customer experiences and reducing waste in their retail operations. Examine the technical hurdles of deploying large object detection models on edge devices and understand the strategies for optimizing computer vision systems that bridge the gap between digital interfaces and physical world applications in robotics, autonomous driving, and automated retail environments.