Biometrics based methods and systems for user authentication

• US 8,942,431 B2
• Filed: 06/24/2012
• Issued: 01/27/2015
• Est. Priority Date: 06/24/2012
• Status: Expired due to Fees

• Current Assignee: Neurologix Security Incorporated
• Original Assignee: Neurologix Security Group Incorporated
• Inventors: Forero, James, Dziuba, Dmitry Alexandrovich
• Primary Examiner(s): Bali, Vikkram
• Application Number:
  US13/531,550
• Publication Number:
  US 20130343616A1
• Time in Patent Office: 947 Days
• Field of Search: None
• US Class Current: 382/115
• 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 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

Disclosed are computer-implemented methods and systems for authentication of users through capturing and analyzing biometrics data such as keystroke dynamics. Once the keystroke dynamics is acquired, a corresponding feature vector related is calculated by calculating multiple sub-feature values using various algorithms such as normalization, relations between keystroke dynamics events, relations of sums of various sub-feature values, and so forth. The feature vector is then selectively compared to a group of reference feature vectors associated with the user. If it is determined that the newly obtained feature vector is similar to the selected group reference feature vectors, the user is authenticated. The process of determining similarity involves decomposition of feature vectors and reference feature vectors onto multiple sub-vectors and calculation similarity therebetween.

Claims

1. A computer-implemented method for user authentication, the method comprising:

receiving an input of a user to be authenticated;

acquiring biometric data of the user;

generating a plurality of sub-feature values based upon the biometric data, wherein each of the sub-feature values is calculated as

2. The method of claim 1, wherein the input comprising a login and/or a password.

3. The method of claim 1, wherein the biometric data comprising one or more of the following:

a typing rhythm, keystroke dynamics, a time period between two or more keys are pressed, a time period spent during one or more keys are pressed, a force applied by the user against a key, a timing pattern, a pattern of user motor skills, or a combination thereof.

4. The method of claim 1, wherein v_i define keystroke dynamics using the following equations:

v₁=t₂−

t₁, v₂=t₃−

t₂, . . . , v_k-1=t_k−

t_k-1

where t₁ . . . t_k represents a sequence of time intervals between a key pressed and depressed for one or more keys.

5. The method of claim 1, wherein the one or more sub-feature values are calculated as follows:

6. The method of claim 1, wherein the one or more sub-feature values are calculated as follows:

7. The method of claim 1, wherein the one or more sub-feature values are calculated as follows:

8. The method of claim 1, wherein determining similarity between the feature vector and the set of reference feature vectors comprising:

breaking the feature vector onto a first set of sub-vectors;

breaking the set of reference feature vectors onto a second set of sub-vectors; and

comparing the first set of sub-vectors and the second set of sub-vectors.

9. The method of claim 8, further comprising:

calculating Euclidian distance between the first set of sub-vectors and the second set of sub-vectors, calculating a statistical average delta vector between the first set of sub-vectors and the second set of sub-vectors, calculating Mahalanobis distance, or a combination thereof.

10. The method of claim 8, further comprising:

calculating one or more similarity values based upon the result of comparison of the first set of sub-vectors and the second set of sub-vectors.

11. The method of claim 10, wherein determining similarity between the feature vector and the set of reference feature vectors comprising comparing the one or more similarity values with a threshold, wherein the threshold is a predetermined tunable parameter.

12. The method of claim 1, wherein determining similarity between the feature vector and the set of reference feature vectors comprising applying one or more machine learning algorithms, wherein the machine learning algorithms comprising one or more of the following:

neural networks based algorithms, Support Vector Machine (SVM) algorithms, and k-nearest neighbor (k-NN) algorithms.

13. The method of claim 12, further comprising:

training the at least one machine learning algorithm each time the input is received and the user is authenticated.

14. The method of claim 1, further comprising:

updating the set of reference feature vectors associated with the user each time the input is received and the user is authenticated.

15. The method of claim 14, further comprising:

selecting the least similar reference feature vector among the reference feature vectors associated with the user; and

replacing the least similar reference feature vector with the feature vector.

16. The method of claim 1, further comprising:

acquiring additional biometrics data of the user, the additional biometrics data comprising behavioral biometrics and/or physical biometric data;

generating an additional feature vector based upon the additional biometric data;

determining similarity between the additional feature vector and the set of reference feature vectors; and

making the authentication decision based upon the determined similarity between the additional feature vector and the set of reference feature vectors.

17. A system for user authentication, the system comprising:

at least one processor and a memory storing processor-executable codes, which processor-executable codes, when executed by the at least one processor, cause to implement a method, comprising;

receiving an input of a user to be authenticated and acquiring biometric data of the user;

generating a plurality of sub-feature values based upon the biometric data, wherein each of the sub-feature values is calculated as

18. A non-transitory processor-readable medium having instructions stored thereon, which when executed by one or more processors, cause the one or more processors to:

receive an input of a user to be authenticated;

acquire biometric data of the user;

generate a plurality of sub-feature values based upon the biometric data, wherein each of the sub-feature values is calculated as