Intelligent video applications built using existing cameras or newer smart cameras, and edge gateways are at the core of a massive wave of innovation benefiting our customers.
According to the IoT Signals report, the vast majority of companies with a video AI solution strategy see it as an integral part of their IoT solution. Yet, the reality is that building such a solution can be complicated. Azure Video Analyzer is an Azure Applied AI service that greatly reduces the effort needed to build intelligent video applications by capturing, analyzing, and publishing video & inference on the edge or the cloud.
"Vision and AI capabilities on edge devices are helping companies create companies create breakthrough applications," said Moe Tanabian, Vice President, and General Manager of Azure Edge Devices, Platform & Services, Microsoft. With the latest updates to Azure Video Analyzer, we are making it easier than ever for our customers to build comprehensive AI-powered solutions with actionable insights from videos”
Figure 1: Azure Video Analyzer overview
New Azure Video Analyzer innovations available as of Microsoft Ignite 2021
Since its introduction earlier this year, Azure Video Analyzer capabilities have grown significantly to meet your needs. The Ignite 2021 (November) release provides you with the following new capabilities and enhancements:
- Cloud-native video management solutions
- Visualization of insights from intelligent video applications
- Windows devices optimized for Video AI
The following illustration provides an overview of both existing Azure Video Analyzer capabilities and the new capabilities made available for Ignite 2021.
Figure 2: Azure Video Analyzer capabilities
1. Cloud-native video management solutions
Video Surveillance as a Service (VSaaS) solutions have the convenience of ease of use coupled with the scalability and manageability of cloud computing, making them attractive to enterprises adopting cloud-native solutions. Such solutions offer the same benefits as other cloud services, such as managed infrastructure for IT teams, easy customization, and integration, and remote access from anywhere. Our customers can take advantage of the following features to achieve the desired outcome in these areas.
- Azure Video Analyzer enables you to build a VSaaS system to record, store, view, and export video clips from cameras. You can connect RTSP cameras directly to the cloud or via a lightweight gateway device for network isolation. There's no need to be on-site to view footage and investigate issues; end users can simply log into a cloud-based dashboard via a web browser or mobile app and view the video stream in near real-time (~2 sec latency).
Figure 3: Camera to Azure Video Analyzer Cloud (with or without gateway)
- Azure Video Analyzer makes it effortless to create actionable events along with video clips that can be shared with relevant end users quickly & securely via the cloud. Customers are always in control of the video by defining access policies to meet regulatory requirements.
- Not every video needs to be processed in real-time at the edge or cloud. With the recent updates to the platform, customers can export a portion of the pre-recorded video stream. This enables scenarios related to long-term storage of relevant snippets, exported clips for AI model training, and video evidence management scenarios.
Figure 4: Azure Video Analyzer low latency video playback
“Safety, security and productivity are essential elements for the growth and sustainability of every society and business. Together Axis Communications and Azure Video Analyzer are empowering developers with the tools to rapidly build and deploy AI-powered solutions that improve operational agility, optimize business efficiency and enhance safety and security” Fredrik Nilsson, Vice President, Americas, Axis Communications
2. Visualization of insights from Intelligent video application
Visualization of AI inference data is necessary to make sense of data generated from Video AI systems. Customers need data visualization to detect anomalies or trends, and so on. With this objective in mind:
- You can use widget player npm (node package manager) to customize and readily embed into Power BI or other business intelligence systems or Power Apps or custom applications.
- It can be time-consuming to obtain coordinates required to build an intelligent video application (i.e., crossing line, zone). The video player widget also provides an easy tool to draw lines and zones.
- This release also marks a significant milestone in our innovations with Azure’s Computer Vision for spatial analysis. Customers can leverage new capabilities like tracking a person's speed and ground orientation and display them with our widget player.
- During upcoming updates, we will enable customers
- To overlay lines/zones for easy visualization while viewing the video from the camera
- To define user attributes (i.e., vests, helmets, bags) and use them with the spatial analysis AI service.
These product investments in Azure Video Analyzer will enable you to visualize actionable insights quickly.
Figure 5: Azure Video Analyzer insights visualization
"We were impressed with the comprehensiveness of the Azure Video Analyzer to build out solutions quickly. The integrations between the Azure Video Analyzer and Scenera's Platform as a Service will help our customers gain valuable business insights using the solution's scene-based event analytics.." Andrew Wajs, CTO & Founder of Scenera
3. Windows devices optimized for Video analytics
Manageability of Windows devices and the necessity to run Linux-based containers are key drivers for our investments with “Edge For Linux On Windows” EFLOW. We now guide our developer ecosystem through the PowerShell experience of running EFLOW with Azure Video Analyzer to strengthen this commitment further.
Figure 6: Azure Video Analyzer + EFLOW powering inventory management solution
Finally, use the following resources to learn more about the overall product and services capabilities.
Please contact the Azure Video Analyzer product team at firstname.lastname@example.org for feedback and deeper engagement discussions.