Media analytics performed on live media or distributed video streams can help service providers, content aggregators, and content delivery networks better understand the nature of the visual content and derive useful intelligence from it. Sophisticated media analytics applications can do everything from detecting suspicious intruders in a video surveillance feed to surveying the traffic patterns in a smart city to better control flow. With billions of pieces of visual content exchanged daily, the market opportunities are expansive for creating new, useful services and adding features and capabilities to existing services.
Examples of Media Analytics Use Cases
- Smart City*—Media analytics applications within the Smart City range from identifying vehicles, including their license plates, for traffic monitoring and toll collection to performing facial recognition of individuals near buildings or public spaces, for both security reasons and for crowd analysis. Often visual recognition tasks need to be performed in real time, working most efficiently in edge computing implementations.
- AI-guided video encoding and live streaming—By adding AI capabilities to video encoding algorithms, broadcasters and OTT producers can control quality and bandwidth parameters to meet specified goals. An AI algorithm at the heart of this process can be trained for optimal quality or bandwidth from a collection of video streams, and then automatically adjust the encoder to ensure goals are achieved.
- Offline media analytics—By using object and performer detection and classification, AI applications can provide a wide variety of services to help cloud service providers (CSP) and communications service providers (CoSP) create new revenue channels. For example, AI techniques can identify and tag potentially violent movies, movies that include certain performers, or provide metadata to recommendation services based on movie content. These kinds of services can be applied at different points in the delivery system, in real time or offline, to add intelligence to the packaging and distribution of content.
- Immersive media enhancement—Applying analytics to immersive media applications opens opportunities in both 360-degree content display and AR applications. For example, sports presentations can benefit from tracking players or the ball, using AI to control the encoding for providing a selective view of the action. Interactive applications include being able to automatically overlay video content with relevant information (such as a player’s batting average or basketball foul shot percentage), or an international football team’s ball possession or passing accuracy statistics.
List of Media Analytics Open Source Ingredients
- Scalable Video Technology - Open source project home
- OpenVINO™ Toolkit
- Intel® Media SDK - Developer home
- FFMPEG - Project home
- GStreamer - Project home
Getting Started with Dockerfiles
To help developers get started with setting up a server image and configuring everything needed, we have created Dockerfiles to build containers that are easy to evaluate on private or public clouds. To get started, visit our Docker home page to learn about Dockerfiles and how to install them. If you just want to jump right in, visit the Media Analytics Dockerfile page.