Video Indexer has been enhanced with new features
Published date: 02 April, 2019
Video Indexer has been enhanced with new features. It now automatically collects manual transcript edits into a new entry in the specific language model you use. Use it to improve the accuracy of an out-of-the-box language model with manual edits over a period of time so it will align with your industry-specific vocabulary.
Another new feature is the ability to segment videos by semantic scenes based on visual cues. Scenes can now be used to group together a set of insights and refer to them as insights of the same context in order to deduce a more complex meaning from them. They can also be used as a unit and extracted as a clip from a longer video. Video Indexer also now has the ability to identify rolling credits at the end of a movie or TV show, which can be used by broadcasters to identify when to recommend a new program before losing the audience.
Finally, Video Indexer is now available in eight regions: West US, West US 2, East US, East US 2, West Asia, Southeast Asia, North Europe, and West Europe.