Biometric authentication technology

• US 9,558,415 B2
• Filed: 02/13/2015
• Issued: 01/31/2017
• Est. Priority Date: 06/07/2011
• Status: Active Grant

• Current Assignee: Accenture Global Services Limited (Accenture PLC)
• Original Assignee: Accenture Global Services Limited (Accenture PLC)
• Inventors: Paul, Sanjoy, Tiwari, Abhinav
• Primary Examiner(s): Sherali, Ishrat I
• Application Number:
  US14/621,819
• Publication Number:
  US 20150161460A1
• Time in Patent Office: 718 Days
• Field of Search: 382/115
• US Class Current: 1/1
• CPC Class Codes: G06F 16/51 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 21/32 using biometric data, e.g. fingerprints, iris scans or voiceprints
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
• CPC Class Codes: G06V 40/50 Maintenance of biometric data or enrolment thereof
• CPC Class Codes: G06V 40/70 Multimodal biometrics, e.g. combining information from different biometric modalities

Abstract

Biometric authentication technology, in which biometric data is maintained for a group of people. The biometric data includes sorted similarity scores, where each of the sorted similarity scores is computed to represent similarity between a corresponding biometric image and a reference image. A biometric verification system accesses a biometric image of at least a portion of a person, accesses the reference image, and computes a similarity score that represents similarity between the accessed biometric image and the reference image. The biometric verification system searches the sorted similarity scores included in the biometric data using the computed similarity score and outputs a result based on the searching.

Claims

1. A biometric authentication system comprising:

a data storage system configured to maintain, for a group of multiple, different people, biometric data that includes sorted similarity scores, wherein each similarity score represents a similarity between a biometric image of at least a portion of a corresponding person from the group of multiple, different people and a reference image; and

a biometric verification system that includes at least one processor, the biometric verification system being configured to;

access data for a particular biometric image of at least a portion of a particular person;

compute a particular similarity score that represents similarity between the particular biometric image and the reference image used in computing the similarity scores for each person in the group of multiple, different people;

search, using the computed particular similarity score, the sorted similarity scores included in the biometric data for the group of multiple, different people to find a closest match to the computed particular similarity score in the sorted similarity scores;

compute a difference between the computed particular similarity score and a similarity score of the closest match;

determine whether the difference satisfies a threshold difference value; and

output the closest match based on a determination that the difference satisfies the threshold difference value.

2. The biometric authentication system of claim 1, wherein the biometric verification system is configured to search the sorted similarity scores included in the biometric data using the computed particular similarity score by searching the sorted similarity scores in a manner that leverages knowledge of how the similarity scores included in the biometric data are sorted.

3. The biometric authentication system of claim 2, wherein the biometric verification system is configured to search the sorted similarity scores in a manner that leverages knowledge of how the similarity scores included in the biometric data are sorted by performing a binary search of the sorted similarity scores using the computed particular similarity score.

4. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to access the data for the particular biometric image of at least the portion of the particular person by accessing data for an image of an iris of the particular person; and

wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by computing a similarity score that represents similarity between the image of the iris of the particular person and the reference image of the iris of the reference person.

5. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to access the data for the particular biometric image of at least the portion of the particular person by accessing data for an image of a fingerprint of the particular person; and

wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by computing a similarity score that represents similarity between the image of the fingerprint of the particular person and the reference image of the fingerprint of the reference person.

6. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to access the data for the particular biometric image of at least the portion of the particular person by accessing data for an image of a face of the particular person; and

wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by computing a similarity score that represents similarity between the image of the face of the particular person and the reference image of the face of the reference person.

7. The biometric authentication system of claim 1 comprising:

a biometric enrollment system configured to determine, for each person in the group of multiple, different people, the similarity score that represents a similarity between a biometric image of at least a portion of the corresponding person from the group of multiple, different people and the reference image and sort biometric data that includes all of the determined similarity scores using the determined similarity scores, wherein the reference image is used in determining all of the similarity scores included in the biometric data.

8. The biometric authentication system of claim 7:

wherein the biometric data maintained by the data storage system comprises the corresponding biometric image or blob data descriptive of the corresponding biometric image; and

wherein the biometric enrollment system is configured to link, for each person in the group of multiple, different people, the similarity score for the corresponding person with the corresponding biometric image or the blob data.

9. The biometric authentication system of claim 8, wherein the data storage system comprises a memory storage system and a disk storage system, and the data storage system is configured to maintain the sorted similarity scores in the memory storage system and maintain the corresponding biometric image or the blob data in the disk storage system.

10. The biometric authentication system of claim 7, wherein the biometric enrollment system is configured to enroll a new person in the data storage system by:

accessing data for a new biometric image of at least a portion of the new person;

computing a new similarity score that represents similarity between the new biometric image and the reference image;

determining a unique identifier for the new person;

storing, in the data storage system and in association with the unique identifier, the new similarity score at an appropriate point in the sorted similarity scores; and

storing, in the data storage system and in association with the unique identifier, the corresponding biometric image or blob data descriptive of the corresponding biometric image.

11. The biometric authentication system of claim 1, wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by:

accessing data for multiple biometric images of the portion of the particular person;

computing, for each of the multiple biometric images, a similarity score that represents similarity between the corresponding biometric image and the reference image; and

computing an average similarity score based on the similarity scores computed for the multiple biometric images.

12. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to search the sorted similarity scores included in the biometric data using the computed particular similarity score by;

using a search process that leverages knowledge of how the similarity scores included in the biometric data are sorted to find a subset of the people that are each associated with a similarity score within a threshold of the computed particular similarity score;

accessing corresponding biometric image or blob data descriptive of the corresponding biometric image for the subset of the people;

comparing the accessed corresponding biometric image or the accessed corresponding blob data for the subset of the people to the particular biometric image or blob data descriptive of the particular biometric image for the particular person; and

based on comparison results, identifying the closest match.

13. The biometric authentication system of claim 1, wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by using an image similarity process to generate a similarity measure between the particular biometric image and the reference image and normalizing the similarity measure to a similarity score between zero and one.

14. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to access the data for the particular biometric image of at least the portion of the particular person by accessing data for multiple biometric images of the particular person, each of the multiple biometric images corresponding to a different portion of the particular person;

wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by;

computing multiple similarity scores, each of the multiple similarity scores representing similarity between one of the multiple biometric images and one of multiple reference images; and

computing a fused similarity score based on the multiple similarity scores; and

wherein the biometric verification system is configured to search the sorted similarity scores included in the biometric data using the computed particular similarity score by searching the sorted similarity scores included in the biometric data using the fused similarity score.

15. The biometric authentication system of claim 14, wherein the biometric verification system is configured to compute the fused similarity score based on the multiple similarity scores by taking a weighted average of the multiple similarity scores.

16. The biometric authentication system of claim 1:

wherein the biometric verification system is configured to access the data for the particular biometric image of at least the portion of the particular person by accessing data for multiple biometric images of the particular person, each of the multiple biometric images corresponding to a different portion of the particular person;

wherein the biometric verification system is configured to compute the particular similarity score that represents similarity between the particular biometric image and the reference image by;

computing multiple similarity scores, each of the multiple similarity scores representing similarity between one of the multiple biometric images and one of multiple reference images; and

arranging the multiple similarity scores in a vector of similarity scores; and

wherein the biometric verification system is configured to search the sorted similarity scores included in the biometric data using the computed particular similarity score by searching the sorted similarity scores included in the biometric data using the vector of similarity scores.

17. The biometric authentication system of claim 1, wherein the biometric verification system is configured to output the closest match based on the determination that the difference satisfies the threshold difference value by outputting the data, from the biometric data, for the particular person to a device for presentation on a display.

18. A method comprising:

maintaining, in electronic storage and for a group of multiple, different people, biometric data that includes sorted similarity scores, wherein each similarity score represents a similarity between a biometric image of at least a portion of a corresponding person from the group of multiple, different people and a reference image;

accessing data for a particular biometric image of at least a portion of a particular person;

computing a particular similarity score that represents similarity between the particular biometric image and the reference image used in computing the similarity scores for each person in the group of multiple, different people;

searching, by a biometric authentication system that includes at least one processor and using the computed particular similarity score, the sorted similarity scores included in the biometric data for the group of multiple, different people to find a closest match to the computed particular similarity score in the sorted similarity scores;

computing a difference between the computed particular similarity score and a similarity score of the closest match;

determining whether the difference satisfies a threshold difference value; and

outputting the closest match based on a determination that the difference satisfies the threshold difference value.

19. At least one non-transitory computer-readable storage medium encoded with executable instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:

maintaining, in electronic storage and for a group of multiple, different people, biometric data that includes sorted similarity scores, wherein each similarity score represents a similarity between a biometric image of at least a portion of a corresponding person from the group of multiple, different people and a reference image;

accessing data for a particular biometric image of at least a portion of a particular person;

computing a particular similarity score that represents similarity between the particular biometric image and the reference image used in computing the similarity scores for each person in the group of multiple, different people;

searching, by a biometric authentication system that includes at least one processor and using the computed particular similarity score, the sorted similarity scores included in the biometric data for the group of multiple, different people to find a closest match to the computed particular similarity score in the sorted similarity scores;

computing a difference between the computed particular similarity score and a similarity score of the closest match;

determining whether the difference satisfies a threshold difference value; and

outputting the closest match based on a determination that the difference satisfies the threshold difference value.