Shaping classification boundaries in template protection systems

1. An authentication apparatus for authentication of a physical object comprising

an enrollment means (ENRL), and

an authentication means (AUTH), wherein

said enrollment means are arranged in

measuring reference characteristics of said physical object;

generating and storing template data based on said measured reference characteristics;

generating and storing a helper data (W 1 ) based on said template data

generating and storing a control value (V 1 );

the authentication means being arranged in:

obtaining a metric (Y) from the physical object to be authenticated;

and further comprising:

a first generation means (GM 1 ) arranged to generate a first property set (C 1 ) using the helper data (W 1 ) and the metric (Y) obtained from the physical object,

a second generation means (GM 2 ) arranged to subject said first property data set (C 1 ) to a noise compensating mapping (NCM) to compensate for measured noise in said metric (Y) thereby generating a second property set (S 1 ),

a comparing means (CMP) arranged to establish a sufficient match between the physical object and the reference object, by verifying whether the second property set (S 1 ) is within a classification boundary (REG 1 ) of the reference object using the first control value (V 1 ) and wherein a noise robust mapping (NRM) determines the classification boundary,

the authentication apparatus characterized in that it further comprises:

a third generation means (GM 3 ) arranged to generate an error measure (ERR) by quantifying the noise removed by the noise compensating mapping (NCM) using the first property set (C 1 ) and the second property set (S 2 ), and

an authentication decision means (ADM) arranged to use said error measure (ERR) to generate an authentication decision (D).

2. An authentication apparatus as claimed in claim 1 , where the third generation means (GM 3 ) is arranged to generate an error measure (ERR) that quantifies the difference between the first property set (C 1 ) and the second property set (S 1 ).

3. An authentication apparatus as claimed in claim 1 , where the third generation means (GM 3 ) arranged to generate an error measure (ERR) comprises:

a fourth generation means (GM 4 ) arranged to generate the third property set (C 2 ) using the noise robust mapping (NRM) on information comprising the second property set (S 1 ), and

a fifth generation means (GM 5 ) arranged to generate the error measure (ERR) by quantifying the difference between the first property set (C 1 ) and the third property set (C 2 ).

4. An authentication apparatus as claimed in claim 1 , where the authentication decision means (ADM) is arranged to use information comprising the error measure (ERR) and the first control value (V 1 ) associated with both the reference object and the helper data (W 1 ), to generate the authentication decision (D) identifying whether the physical object corresponds with the reference object.

5. An authentication apparatus as claimed in claim 1 , where the authentication decision means (ADM) is arranged to use information comprising the error measure (ERR) and the first control value (V 1 ) associated with both the reference object and the helper data (W 1 ), to generate a probability measure (P) related to the authentication decision (D).

6. An authentication apparatus for identification of a physical object using both a helper data (W 1 ) and a control value (V 1 ) associated with a reference object, the apparatus according to claim 1 , further comprising an identity establishing means arranged to establish the identity of the physical object as being identical to that of the reference object.

7. A computer controlled method suitable of authentication of a physical object in a system employing template data protection,

comprising an enrollment process consisting of the steps of:

measuring reference characteristics of the physical object;

generating template data based on said measured reference characteristics;

generating a helper data (W 1 ) based on said template data

generating a control value (V 1 );

the method further including an authentication process consisting of the steps of

obtaining a metric (Y) from the physical object to be authenticated;

generating a first property data set (C 1 ) based on the helper data (W 1 ) and the metric (Y),

subjecting said first property data set (C 1 ) to a noise compensating mapping (NCM) to compensate for measured noise in said metric (Y) thereby;

generating a second property data set (S 1 ),

for authenticating purposes of the physical object verifying whether the second property data set (S 1 ) is within a classification boundary (REG 1 ) of the reference object using the first control value (V 1 ) and wherein a noise robust mapping (NRM) determines the classification boundary,

the method characterized in that it further comprises the following steps:

generating an error measure (ERR) by quantifying the noise measured in said metric (Y) by the noise compensating mapping (NCM) using the first property set (C 1 ) and the second property set (S 2 ), and

generating an authentication decision (D) using said error measure (ERR).

8. A computer controlled method as claimed in 7 , where the step for generating the error measure (ERR) comprises quantifying the difference between the first property set (C 1 ) and the second property set (S 1 ).

9. A computer controlled method as claimed in 7 , where the step for generating the error measure (ERR) comprises a step that:

generates a third property set (C 2 ) using the noise robust mapping (NRM) on information comprising the second property set (S 1 ),

generates the error measure (ERR) by quantifying the difference between the first property set (C 1 ) and the third property set (C 2 ).

10. A computer controlled method as claimed in claim 7 , where the authentication decision is generated using information comprising the error measure (ERR) and a control value (V 1 ) associated with both the reference object and the helper data (W 1 ), to generate the authentication decision (D) identifying whether the physical object corresponds with the reference object.

11. A computer controlled method as claimed in claim 7 , where the authentication method generates an interim decision (ID) by comparing information derived from the second property set (S 1 ) with information derived from the control value (V 1 ).

12. A computer controlled method as claimed in claim 11 , where the authentication method comprises a step for comparing the error measure (ERR) with a predetermined threshold (t), and where information comprising the outcome of said comparison is used to accept or reject the interim decision (ID), resulting in a authentication decision (D) that is identical to, or different from the interim decision (ID) respectively.

13. A computer controlled method as claimed in claim 7 , where the authentication method uses information comprising the error measure (ERR) and the first control value (V 1 ) associated with both the reference object and the helper data (W 1 ), to generate a probability measure (P) related to the authentication decision (D).

14. A computer controlled method as claimed in claim 7 , where said noise compensating mapping (NCM) is an error correcting code decoder algorithm.

15. A computer controlled method as claimed in claim 7 , further comprising a step to acquire an alleged identity.

16. A computer controlled method as claimed in claim 7 , further comprising a step for generating a second control value (V 2 ) by applying a one way function h on the second property set (S 1 ), where said second control value (V 2 ) is used to establish a sufficient match between the physical object and the reference object.

17. A computer controlled method of identification of a physical object using both a helper data (W 1 ) and a control value (V 1 ) associated with a reference object, the method according to claim 7 , further comprising a step for establishing the identity of the physical object as being identical to that of the reference object.

18. A computer controlled method as claimed in claim 17 , where the identification method uses information comprising the error measure (ERR) and the first control value (V 1 ) associated with both the reference object and the helper data (W 1 ), to generate a probability measure (P).

19. A computer controlled method as claimed in 18 , where said probability measure (P) is used to establish the identity of said physical object.

20. A computer program product comprising program code means stored on a computer readable medium for performing the method of claim 7 when said program product is run on a computer.