Searching YouTube for a specific video can be incredibly frustrating – you're relying on someone entering the right keywords, not making any spelling mistakes or simply not lying about what the video features.
Videosurf claims to change all that. It's a new computer vision and search engine that "sees inside" videos and analyses its actual content, even to the point of identifying people in the clip.
Multigrid fast computation, you know
Videosurf combines 'multigrid fast computation' (that's 'poncey maths' to you and me) and serious computing muscle to detect scene breaks, spot key frames (and key actors) and index the clip automatically.
Search results are presented as a series of thumbnails from the video, selected by scene, so that you can check the video is the one you're after and even skip straight to the interesting bits.
VideoSurf aggregates content from major video websites, including YouTube and Hulu, as well as major TV networks like Comedy Central and ESPN, although the engine is only available as a closed beta at the moment.
It's all about advertising...
Think all this technology has been made to help you find that episode of The Goodies starring John Cleese? Think again.
"Television networks and advertisers are eagerly awaiting new search techniques that will more effectively connect consumers to online video content," said Joel Hyatt, chairman of the board for VideoSurf. "VideoSurf has cracked the code on video search, and created a significant new revenue opportunity for content owners and advertisers."
Get daily insight, inspiration and deals in your inbox
Get the hottest deals available in your inbox plus news, reviews, opinion, analysis and more from the TechRadar team.
Mark Harris is Senior Research Director at Gartner.