|A patent filing recently made available from Google indicates facial recognition for YouTube videos will soon be implemented. The technology also has implications for Google Project Glass. Having trouble putting a face to a name may be a thing of the past if the technique is used on the upcoming goggles.|
The technique involves creating facial tracks from the video inputs and selecting the key images for each track. Then creating a cluster of face tracks to form face cluster and then creating face models from these face clusters.
The feature allows Google to attach name to people in videos even at various angles and different light conditions.
Quite possibly when Google releases this, people will be able to tag others in the videos.
Also, when Google’s project Glass lands in the markets in 2014, the facial recognition method for video is going to make Google Glass augment the users heads-up display by updating information about faces encountered while wearing the device. This was conceptualized in the Project Glass demo video released by Google in April.
SOURCE Crazy Engineers
|By 33rd Square||Subscribe to 33rd Square|