Demo: Video Analytics – Killer App for Edge Computing
The world is witnessing an unprecedented increase in camera deployment. The USA and UK, for instance, have one camera for every 8 people. Video analytics from these cameras are becoming more and more pervasive, exerting important functions on a wide range of verticals including manufacturing, transportation, and retails. While vision techniques have seen considerable advancement, they have come at the expense of compute and network cost.
As an alternative to the centralized, in-the-cloud compute paradigm, edge computing offers the promise of near real-time insights, faster localized actions, and cost reduction because of efficient data management and operations. We believe video analytics may represent the “killer application” for edge computing due to its demanding requirements on compute, bandwidth and latency. In this demo, we showcase a live video analytics system that spans across the cloud and edge, with low cost, and produces results with high accuracy.
Today, video cameras are being used at a large scale by public and private enterprises for a variety of reasons—from security surveillance and traffic planning to consumer support in retail and hospitality settings. Thanks to gains in computer vision, particularly object detection and classification, video analysis has become more accurate. Fast and affordable real-time analysis, however, is lagging. Project Rocket seeks to make easy, cost-effective video analysis of live camera streams a reality. Project Rocket, an extensible software stack that leverages the edge and cloud, is designed with maximum functionality in mind, capable of meeting the needs of varying video analytic applications. In this webinar, Microsoft researchers Ganesh Ananthanarayanan and Yuanchao Shu explain how Rocket—now open source on GitHub—uses approximation to run scalable analytics across the edge and cloud and how efficient live video analysis advances the interactive querying of stored video. The researchers will also provide a tutorial on how to get started with the stack and how to construct and execute video analytics pipelines. Together, you'll explore: Exciting applications of video analytics Techniques that make continuous video analytics dramatically cheaper Key components of the Rocket video analytics software stack Building your own video analytics applications Resource list: Microsoft Rocket (Project page) Microsoft Rocket (GitHub) Live video analytics and research as Test Cricket with Dr. Ganesh Ananthanarayanan (Podcast) Cracking open the DNN black-box (Publication) Networked Cameras Are the New Big Data Clusters (Publication) Demo: Video Analytics - Killer App for Edge Computing (Publication) Ganesh Ananthanarayanan (Researcher profile) Yuanchao Shu (Researcher profile) *This on-demand webinar features a previously recorded Q&A session and open captioning. This webinar originally aired on December 12, 2019 Explore more Microsoft Research webinars: https://aka.ms/msrwebinars