Viewdle is a facial recognition powered video seach engine for indexing, searching and recognizing actual appearances of people inside a video in real-time. Viewdle’s facial-recognition visual analysis technology automatically looks “inside a video” to identify true on-screen public appearance with unrivaled accuracy and relevance.

Viewdle’s frame-by-frame analysis provides search results based upon the actual video asset, not just the associated metadata. When you search for Bill Gates in a Viewdle powered search engine, the service will do a frame by frame analysis of all videos in the index and compare it with the face of Bill Gates stored in its own facial recognition database. If any video matches, the result is displayed. This is unlike other video search engines like Google Video, DailyMotion or Metacafe where results are returned based on individual tags (metadata) submitted by the users themselves, even though the said video may not contain an appearance of the searched personality.

Another useful feature is that you can directly jump to the part of the video part where the personality is featured.

This service made its debut today at Techcrunch40 and is available for testing on Reuter’s Face Search website.