Authentication systems and authentication method

• US 7,913,090 B2
• Filed: 08/12/2005
• Issued: 03/22/2011
• Est. Priority Date: 08/19/2004
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Abe, Hiroshi
• Primary Examiner(s): Song; Hosuk
• Application Number:
  US11/573,859
• Publication Number:
  US 20080040614A1
• Time in Patent Office: 2,048 Days
• Field of Search: 713182-186, 713/168, 726 2- 5, 382115-118, 382/124
• US Class Current: 713/186
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/2415 based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Abstract

An authentication system and method thereof are provided. A threshold value determination unit determines a threshold value Xth so that a Mahalanobis distance defined by a mean value μt of a true profile and a standard deviation σt and the Mahalanobis distance defined by a mean value μo of the other profile and a standard deviation σo coincide.

Claims

1. An authentication system for comparing a correlation value between inspected characteristic data of biometric data acquired from a biometric subject being inspected and comparative data linked with the predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

a processor; and

a memory device storing instructions which when executed by the processor, cause the processor to;

store a threshold value defined so that a value obtained by dividing an absolute value of a difference between a first mean value of a plurality of first correlation values generated by detecting correlation with said comparative data for a plurality of first characteristic data previously acquired from said predetermined biometric subject and said threshold value by a standard deviation of said plurality of first correlation values and values obtained by dividing an absolute value of a difference between a second mean value of a plurality of second correlation values generated by detecting correlation with said comparative data for a plurality of second characteristic data previously acquired from a biometric subject other than said predetermined biometric subject and said threshold value by said plurality of second correlation values coincide; and

compare said inspected characteristic data and the threshold value to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject.

2. An authentication system as set forth in claim 1, wherein the instructions, when executed by the processor, cause the processor to determine a threshold value based on a first normal profile defined by said plurality of first correlation values and a second normal profile defined by said plurality of second correlation values.

3. An authentication system as set forth in claim 2, wherein said threshold value is based on a third normal profile defined linearly symmetrically using a pattern positioned at said first normal profile side with respect to a maximum frequency in said second normal profile.

4. An authentication method comparing a correlation value between inspected characteristic data of a biometric subject acquired from the biometric subject being inspected and comparative data linked with a predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

a first step of determining a threshold value so that a value obtained by dividing an absolute value of a difference between a first mean value of a plurality of first correlation values generated by detecting correlation with said comparative data for a plurality of first characteristic data previously acquired from said predetermined biometric subject and said threshold value by a standard deviation of said plurality of first correlation values and values obtained by dividing an absolute value of a difference between a second mean value of a plurality of second correlation values generated by detecting correlation with said comparative data for a plurality of second characteristic data previously acquired from a biometric subject other than said predetermined biometric subject and said threshold value by said plurality of second correlation values coincide; and

a second step of authenticating whether or not said biometric subject being inspected is said predetermined biometric subject by comparing said inspected characteristic data and said threshold value determined at said first step.

5. An authentication method as set forth in claim 4, wherein said first step determines said threshold value based on a third normal profile defined linearly symmetrically using a pattern positioned at said first normal profile side with respect to a maximum frequency in said second normal profile when said second normal profile defined by said plurality of second characteristic data is non-linearly symmetric.

6. An authentication system comparing a correlation value between inspected characteristic data of a biometric subject acquired from the biometric subject being inspected and comparative data linked with a predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

an inputting means for inputting a false rejection rate showing a probability of certifying that said biometric subject being inspected who is said predetermined biometric subject is not said predetermined biometric subject;

a processor; and

a memory device storing instructions which when executed by the processor, cause the processor to;

determine said threshold value so as to satisfy said false rejection rate input by said inputting means by assuming that a plurality of correlation values generated by detecting correlation with said comparative data for the plurality of characteristic data previously acquired from said predetermined biometric subject are according to a normal profile; and

authenticate whether or not said biometric subject being inspected is said predetermined biometric subject by comparing the correlation value of said inspected characteristic data and said comparative data and said threshold value.

7. An authentication system as set forth in claim 6, wherein the instructions, when executed by the processor, cause the processor to store correspondence data showing linked together a Mahalanobis distance dt defined for each of a plurality of values separated by a predetermined value from a mean value of said plurality of correlation values and showing that value normalized by a standard deviation of said plurality of correlation values and said false rejection rate, and

wherein said Mahalanobis distance dt corresponding to said false rejection rate input by said inputting means is based on said correspondence data and said threshold value is based on said specified Mahalanobis distance dt, said mean value, and said standard deviation.

8. An authentication method comparing a correlation value between inspected characteristic data of a biometric subject acquired from the biometric subject being inspected and comparative data linked with a predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

a first step of inputting a false rejection rate showing the probability of certifying that said biometric subject being inspected which is said predetermined biometric subject is not said predetermined biometric subject;

a second step of determining said threshold value so as to satisfy said false rejection rate input in said first step by assuming that a plurality of correlation values generated by detecting correlation with said comparative data for the plurality of characteristic data previously acquired from said predetermined biometric subject are according to a normal profile; and

a third step of authenticating whether or not said biometric subject being inspected is said predetermined biometric subject by comparing the correlation value of said inspected characteristic data and said comparative data and said threshold value determined in said second step.

9. An authentication system comparing a correlation value between inspected characteristic data of a biometric subject acquired from the biometric subject being inspected and comparative data linked with a predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

an inputting means for inputting a false acceptance rate showing the probability of certifying that said biometric subject being inspected who is not said predetermined biometric subject is said predetermined biometric subject;

a processor; and

a memory device storing instructions which when executed by the processor, cause the processor to;

determine said threshold value so as to satisfy said false acceptance rate input by said inputting means by assuming that a plurality of correlation values generated by detecting correlation with said comparative data for the plurality of characteristic data previously acquired from said predetermined biometric subject are according to a normal profile; and

authenticate whether or not said biometric subject being inspected is said predetermined biometric subject by comparing the correlation value of said inspected characteristic data and said comparative data and said threshold value.

10. An authentication system as set forth in claim 9, wherein the instructions, when executed by the processor, cause the processor to store correspondence data showing linked together a Mahalanobis distance dt defined for each of a plurality of values separated by a predetermined value from a mean value of said plurality of correlation values and showing that value normalized by a standard deviation of said plurality of correlation values and said false acceptance rate, and

wherein said Mahalanobis distance do corresponding to said false acceptance rate input by said inputting means is based on said correspondence data and said threshold value is based on said specified Mahalanobis distance do, said mean value, and said standard deviation.

11. An authentication method for comparing a correlation value between inspected characteristic data of a biometric subject acquired from the biometric subject being inspected and comparative data linked with a predetermined biometric subject and a predetermined threshold value linked with said predetermined biometric subject to authenticate whether or not said biometric subject being inspected is said predetermined biometric subject, comprising:

a first step of inputting a false acceptance rate showing the probability of certifying that said biometric subject being inspected which is not said predetermined biometric subject is said predetermined biometric subject;

a second step of determining said threshold value so as to satisfy said false acceptance rate input in said first step by assuming that a plurality of characteristic values generated by detecting correlation with said comparative data for the plurality of characteristic data previously acquired from said predetermined biometric subject are according to a normal profile; and

a third step of authenticating whether or not said biometric subject being inspected is said predetermined biometric subject by comparing the correlation values of said inspected characteristic data and said comparative data and said threshold value determined in said second step.