Metadata is the crux of video. These shorthand textual descriptions can be used for more detail searching, classification or other organizing. But when it comes to scale, machine-generated meta data has issues with accuracy.
But with such a huge volume of video on the web, it makes human validation prohibitively expensive. Because of this, video searching misses a lot of what’s visually there.
But in order to solve this issue, Singapore-based organization called Aivon launched a blockchain-based open protocol that combines AI with human crowdsourcing to support the creation and validation of text and metadata for a video search engine, translation, transcription and brand-safety applications. The organization says the resulting community will create the first decentralized video search engine.