Media Asset Management customers can auto-generate metadata such as keywords, product placement and entity recognition for their content. A set of topics can be deduced from the content and returned as a list.

Take a look at the solutions brief below to learn more.

Read the solutions brief

The Proof of Concept being shown in the demo below demonstrates the extraction of keywords and topics from text on a web page. The content could equally come from a speech-to-text transcript. In this demo, we show text being copied from a web page into a plain text file to create a transcript of a news item. We then run the topic extractor on this news item. The text is processed by Machine Learning algorithms that output two lists: keywords and topics. The keywords are important content words that are located in the news item; topics are summary terms that are not in the news item but are inferred by the Machine Learning algorithm.

Such a topic generator is useful for Media Asset Management (MAM) use cases: indexes and digests can be produced automatically from digital assets. This is useful for recommendation and search capabilities, or any workflow where it is important to automatically extract metadata from spoken or written language in a digital asset.

We’d love to hear whether the Speechmatics Metadata Generation PoC would be useful to you and your business. We’ve created a short form which we’d be grateful if you could fill out. Your feedback is really useful to us in helping to shape the future of the industry and it shouldn’t take any longer than 5 minutes to complete.