Visual Crawling
Yesterday, AOL announced they had purchased Truveo, a two-year-old video search startup company. While AOL is not releasing it’s price value, the word on the net is that the price was $50 million. Not bad for a company that was started in January 2004. But it doesn’t matter how much money this company made, it’s what they actually made to get there.
Truveo is a video search website. They read “meta-data” about a video, similiar to how Flash 8 now includes meta-data. They do not search the closed caption transcripts, but do search the blurbs and rss feeds about the video and relate keywords to that video.
From the Truveo Website:
We felt that if we could build a crawler that could identify the visual characteristics of a typical web application in the same way that a person could, then it would be possible to find and index all of the web video that other crawlers miss. We call this new approach “visual crawling”.
Here’s the video search for Comedy Central’s The Daily Show.
While it seems interesting, I would think to go one more and invent an auto-transcribing software. If you can pull the transcript, even only from a closed-caption transcript, and rid of the noise words like “the”, “next up” & “after these commercials”, then researching mentions of “Manitoba+pollution” or “alex+reid+winnipeg” within video content would be easier.. or rather, it would be possible for the first time.
Producing media is getting cheaper, video cameras and computers are costing less and less every season. Distributing media is incredibly easy with the Internet, using BitTorrent or a simple website. But being found by the masses who don’t have the patience for research is the next step.
No Comments
RSS feed for comments on this post. TrackBack URL
-
Related Posts by Alex Reid:
- » Understanding SOPA & PIPA
- » Where Startups Get Their Money
- » RIP Napster
- » English Language
- » What Would Jello Do
