×

Method and system for facial recognition for a videoconference

  • US 9,282,284 B2
  • Filed: 05/20/2013
  • Issued: 03/08/2016
  • Est. Priority Date: 05/20/2013
  • Status: Active Grant
First Claim
Patent Images

1. A method comprising:

  • identifying a participant from audio information, wherein identifying the participant from the audio information comprises;

    performing a feature extraction and a speaker segmentation on the audio information to determine a voice model, andcomparing the determined voice model from the audio information with a plurality of voice models stored in a database to identify the participant;

    identifying the participant in video information, wherein identifying the participant in the video information comprises;

    identifying a plurality of facial images in the video information;

    determining a one of the plurality of facial images in the video information as having the most movement as compared to others of the plurality of facial images; and

    identifying the participant as the determined one of the plurality of facial images;

    capturing, from the video information, a plurality of images of the participant identified in the video information wherein ones of the plurality of captured images of the participant include respective different expressions of a face of the participant and wherein other ones of the plurality of captured images of the participant include respective different illumination conditions of the face of the participant;

    associating a unique identifier with the captured plurality of images, the unique identifier corresponding to the participant identified from the audio information; and

    saving the captured plurality of images and the associated unique identifier in the database.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×