Method and apparatus employing audio and video data from an individual for authentication purposes

• US 5,761,329 A
• Filed: 12/15/1995
• Issued: 06/02/1998
• Est. Priority Date: 12/15/1995
• Status: Expired due to Term

• Current Assignee: AT&T Corporation (AT&T, Inc.)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Civanlar, Mehmet Reha, Chen, Tsuhan
• Primary Examiner(s): Mancuso, Joseph
• Application Number:
  US08/572,982
• Time in Patent Office: 900 Days
• Field of Search: 382/115, 382/116, 382/118, 340/825.3, 340/825.34
• US Class Current: 382/116
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition
• CPC Class Codes: G06V 10/811 the classifiers operating on different input data, e.g. multi-modal recognition
• CPC Class Codes: G06V 40/70 Multimodal biometrics, e.g. combining information from different biometric modalities
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition

Abstract

A method and apparatus is provided for determining the authenticity of an individual. In accordance with the method, audio and video data of the individual speaking at least one selected phrase is obtained. Identifying audio features and video features are then extracted from the audio data and the video data, respectively. A feature vector is formed which incorporates both the audio features and the video features. The feature vector is compared to a stored feature vector of a validated user speaking the same selected phrase. The individual is authenticated if the feature vector and the stored feature vector form a match within a prescribed threshold.

Claims

1. A method for determining authenticity of an individual, said method comprising the steps of:

obtaining audio data of the individual speaking at least one selected phrase;

obtaining video data of the individual speaking said at least one selected phrase;

extracting identifying audio features and video features from said audio data and said video data, respectively, said audio and video features including temporal variations in said audio data and video data, respectively;

forming a feature vector that incorporates said audio features and said video features;

comparing said feature vector to a stored feature vector of a validated user speaking said at least one selected phrase; and

authenticating said individual if said feature vector and said stored feature vector form a match within a presecribed threshold.

2. The method of claim 1 further comprising the step of prompting the individual to speak a predetermined password.

3. The method of claim 2 wherein said password is selected by the individual.

4. The method of claim 1 wherein the step of extracting identifying audio features comprises the step of extracting variations in distance among a plurality of points on lips of the individual over a period of time.

5. The method of claim 4 wherein said plurality of points include two points on corners of the lips, two points on an upper lip and two points on a lower lip.

6. The method of claim 1 wherein the step of extracting identifying video features employs a mouth model.

7. The method of claim 6 wherein the step of extracting identifying audio features employs LPC cepstrum coefficients.

8. The method of claim 7 wherein the step of comparing said feature vector to a stored feature vector employs time-warped mapping.

9. The method of claim 6 wherein said mouth model employs model-based audiovisual coding.

10. The method of claim 1 wherein the step of extracting identifying audio features employs LPC cepstrum coefficients.

11. The method of claim 1 wherein the step of comparing said feature vector to a stored feature vector employs time-warped mapping.

12. The method of claim 1 wherein the step of obtaining video data includes the step of displaying a previously stored facial template.

13. The method of 12 wherein said monitor displays a mirror image of the facial template.

14. An apparatus for determining authenticity of an individual, comprising:

means for obtaining audio data and video data of the individual speaking at least one selected phrase;

means for extracting identifying audio features and video features from said audio data and said video data, respectively, said audio and video features including temporal variations in said audio data and video data, respectively;

means for forming a feature vector that incorporates said audio features and said video features; and

means for comparing said feature vector to a stored feature vector of a validated user speaking said at least one selected phrase to determine if a match is formed above a prescribed threshold.

15. The apparatus of claim 14 wherein said means for obtaining audio data and video data is a camera and microphone.

16. The apparatus of claim 14 wherein said forming means and said comparing means is geographically remote from said extracting means.

17. The apparatus of claim 16 wherein said extracting means is a data input terminal.

18. The apparatus of claim 16 wherein said extracting means is a client and said forming means and said comparing means is a server in communication with said client.

19. The apparatus of claim 14 wherein said forming means and comparing means is geographically co-located with said extracting means.

20. The apparatus of claim 19 wherein said forming means, said comparing means and said extracting means employ a common processor.

21. The apparatus of claim 14 wherein said means for obtaining audio and video data includes a monitor on which the individual is observable.

22. The apparatus of claim 21 wherein said monitor displays a previously stored facial template.

23. The apparatus of claim 22 wherein said monitor displays a mirror image of the facial template.