AI and ML are now at the core of a growing number of products and tools in the broadcast and media industries. Anne Morris rounds up some of the latest launches which are leveraging these technologies.
Initially focused on automating repetitive tasks, the integration of Artificial intelligence (AI) and machine learning (ML) is opening the door to many other useful applications. IBC365 takes a look at ten recent product launches and service evolutions, some of which have focused on solving the challenges brought about by the coronavirus health crisis.
1. AI-based ad insertion technology
UK-based Mirriad began using AI technology to digitally insert advertisements and products into movies and TV shows after they have been filmed. As reported by IBC365, Mirriad can digitally embed a branded bottle on a table, a new advertisement on an existing billboard, or a commercial running on the TV in the background. The company’s platform uses AI to identify placement opportunities, and then employs visual effects technology to insert real-world objects that weren’t in the original shoot, or overlays existing brand imagery with new product shots. The idea has already generated attention among broadcasters and advertisers. Indeed, producers have been looking for ways to generate additional revenue from their back catalogues after production stalled during the COVID-19 pandemic.
2. AI in video codecs for optimising video flows
IBC365 also reported on how future advances on optimising video streaming workflows will be made in software that is automated by AI. Companies such as Haivision, Harmonic, InterDigital, iSize Technologies, and V-Nova are working on different ways of applying AI and ML techniques within video codecs.
3. Using AI/ML to convert horizontal to vertical formats for smartphones
French news channel BFMTV launched ‘live vertical’ technology that automatically converts the horizontal frames of standard television streams to a vertical format that is better suited to smartphones. Altice-owned BFMTV collaborated with French start-up Wildmoka to develop the product, which allows the traditional horizontal format of television to be automatically ‘rezoned’ into a mobile-friendly vertical format using AI and ML techniques. Wildmoka calls its product Auto ReZone, and designed it to provide a mobile-first, vertical viewing experience for news. The product extracts content from live streams or recorded videos, and automatically reconstitutes it into a mobile-first video format (typically 9:16 or 1:1). The tool uses AI and ML to detect all the zones of interest in each 16:9 frames; select a vertical layout/template suitable to fit the various zones of interest detected above; extract each zone from the horizontal frame and adjust them individually to fit the zone sizes of the target vertical layout; and re-compose the extracted zones and graphical elements into the overall final vertical frame.
4. AI in full RDK video-based product
SoftAtHome launched a new video-based product based on the Reference Design Kit (RDK) open source software for the video industry, and which uses AI techniques for user-friendly navigation and personal data security. The product integrates multicast, DVB, live DASH streaming, a universal search aggregator, new premium video streaming apps, voice controls, and SoftAtHome’s white-label ImpressioTV user interface. The company uses AI algorithms to help optimise the user experience and propose personalised content while keeping data private. In addition, AI-based voice control assets have been integrated into the RDK product to make navigation on TV screens more user-friendly.
5. Using AI and ML to prevent customer churn
Qligent introduced Foresight as a cloud-based service that uses AI, ML, and big data to mitigate content distribution issues, prevent churn, and protect service provider revenue. Foresight is designed to help broadcasters, MVPDs and OTT service providers understand and correlate factors that contribute to higher audience engagement by providing real-time data analytics based on system performance and user behaviour. The aim is to stop so-called silent sufferers from cancelling their subscriptions by predicting and preventing customer churn. AI and ML provide automated data collection, while deep learning technology mines data from hundreds or thousands of layers of data. Big data technology then correlates and aggregates the data for real-time, cloud-based quality assurance, helping service providers to quickly address distribution issues.
6. AI for managing the connected home
Amdocs launched doxi HomeOS, an AI-based cloud-native home operating system (OS) designed to enable service providers to move beyond basic connectivity services in the connected home. doxi HomeOS provides AI-based insights, simple voice commands and touch-free care capabilities to resolve customer support needs. The OS also offers enhanced cybersecurity monitoring capabilities and parental control over the growing number and usage of connected devices and apps in the home. Furthermore, doxi HomeOS offers consumers the ability to self-manage connectivity and WiFi settings as well as automated, AI-based notifications related to usage patterns and media and gaming consumption. Gil Rosen, general manager of amdocs:next, said doxi HomeOS is relevant for all broadband providers, ranging from incumbents looking to differentiate and grow services to CSPs rolling out 5G fixed wireless access to enhance home broadband connectivity.
7. AI in live transcription services
Epiphan Video launched LiveScrypt as a live transcription service. LiveScrypt is a cloud and AI-based speech-to-text transcription service that enables audiences to engage with live events as they happen – regardless of any hearing impediment, native language, or distraction. LiveScrypt is said to transcribe with at least 85% to 90% accuracy. It also adds punctuation and on-the-fly corrections based on the confidence of words in context. North American Industry Classification System (NAICS) codes for standardised industry-related terms are also supported for greater accuracy.
8. AI in robotic cameras
Telemetrics introduced AI techniques as well as motion tracking and servo-mechanical excellence as standard on its latest robotic camera products and systems. For example, the OmniGlide robotic roving platform was improved with new ML algorithms, allowing its shot recall settings to intelligently find the best path within the space in which it is operating. This Path Planning means it can figure out the safest way between point A and point B, even when there’s an obstruction (like a news desk) in between. This is accomplished in tandem with the Telemetrics RCCP-2A robotics and camera control panel running STS software.
9. AI for video-on-demand (VoD)
SPI/ Film Box launched an AI-based content streaming service called FilmBox Plus. The multi-platform service merges linear and on-demand experiences through AI-supported linear channels and video-on-demand (VoD) content. SPI said the new service is an evolution of FilmBox Live, which has been active for over a decade. It is expected that FilmBox Plus will launch globally by the end of 2020, replacing FilmBox Live.
10. Corporate broadcasting using AI-based vPilot
Mobile Viewpoint partnered with BuckDesign to provide broadcast studio services, with a specific emphasis on companies looking to enhance their corporate communications and marketing initiatives. BuckDesign is using vPilot technology with AI automation from Mobile Viewpoint as part of its inhouse broadcast studio, which can be rented by companies wishing to deploy their own professional TV broadcast studio. vPilot is an automated studio system that controls multiple cameras without the need for an onsite director or camera operators. BuckDesign, in conjunction with vPilot, built its own studio in Alkmaar in North Holland that is available to corporates wishing to undertake a production without investing in a complete studio. For companies that wish to implement their own studio, BuckDesign can provide the full set-up from room design to implementation of the technology.
No comments yet