Your browser is not supported. Please update it.

21 September 2005

Truveo visual crawling opens up more video

We’ve gone on and on about video search engines, and largely they are a bit clunky and don’t find what you’re looking for. But a new search engine that has come out of two years of stealth work, called Truveo has impressed us this week.

The company says that web crawlers can’t find most of the video on the web because much of the video out there on the web does not have the right metadata or closed captioning that TV quality video usually has.

Truveo says it takes a fundamentally new approach it calls ‘visual crawling’ which it doesn’t explain, but it says that its crawler behaves more like a person who is viewing each web page with video on, and the proof is that it not only comes up with more video than any other search engine we have seen, but it always manages to play it first time every time, which isn’t the case with any of the others.

For video to be searchable, says Truveo, it is also necessary to collect meaningful text metadata to associate with each video file. It relies like everyone else on mining closed-caption transcripts and importing RSS feeds. But where these are not found the Truveo crawler can ‘visually’ examine the context of the surrounding web application and find detailed metadata related to every video.