GUTERSLOH, Germany—Arvato Systems and its Vidispine business have added advanced cognitive functions to Vidinet that enable its users to execute multiple media analyzers and pre-trained AI models via a single interface to unify, aggregate and normalize results.
The functions are available to users of the Arvato Systems Video Production Management Suite (VPMS) and Broadcast Management Suite (BMS), as well as all Vidispine API-as-a-Service subscribers, the company said.
Vidinet’s unified data model, also known as the Analyzed Data Unit (ADU), is different from simple AI aggregators in the flexibility it brings to cognitive functionality, the company said.
Cascading and combining different services to gain further insight into media is supported in Vidinet. One example of how to take advantage of this flexibility is creating metadata relating to the context and tone of content. Using a speech-to-text service can create a transcript of a clip, which can then be passed through a natural language processing algorithm for tone and context. Leveraging an object recognition analysis increases the level of confidence about that data, it said.
The company plans to unveil other functionality later in the year. Vidinet will provide a common training interface for supported services, making training easily integrated into media workflows.
“This addition to Vidinet takes an existing concept of aggregating AI services and extends it in a way that makes the implementation of AI and ML in media workflows far simpler and more maintainable,” said Erik Ahlin, CEO of Vidispine.
The holistic approach behind the development of the technology and applying it to the Arvato and Vidispine portfolio will enable media companies to respond quickly to changing demands to meet the demands of consumers and advertisers with greater efficiency and profits, added Holger Noske, VP Business Solutions at Arvato Systems.