High-resolution Multi-camera Analysis Infrastructure to Support Future Smart Cities

In the IOWN (Innovative Optical and Wireless Network) era, smart cities will be built on a cyber-physical system (CPS) that makes all information in the city valuable and accessible. In this article, we introduce an artificial intelligence (AI) inference infrastructure that can efficiently process high-resolution, multi-camera images to support an urban-scale CPS. This infrastructure is based on the concept of event-driven inferencing to significantly reduce the processing loads and energy consumption of AI-inference processing executed with the infrastructure. The basic techniques of model cascading and inference resource sharing are discussed in the article. © 2022 Nippon Telegraph and Telephone Corp.. All rights reserved.