Threshold Score Validation

• US 20090003663A1
• Filed: 10/05/2006
• Published: 01/01/2009
• Est. Priority Date: 10/08/2005
• Status: Active Grant

• Current Assignee: Rolls-Royce PLC (Rolls-Royce Holdings Plc)
• Original Assignee: Rolls-Royce PLC (Rolls-Royce Holdings Plc)
• Inventors: Webster, John R.
• Granted Patent: US 8,229,182 B2
• Time in Patent Office: Days
• US Class Current: 382/119
• 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: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• 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 40/30 Writer recognition; Reading and verifying signatures

Abstract

A method is provided whereby confidence in a validation parameter set is achieved by comparison with a known sample of a parameter in terms of the expected template comparison score as well as mean values and deviation from that score. Thus, with respect to user individual identification, biometric parameters such as fingerprint or handwriting or voice recognition can be utilised to compare the stored template for that individual with the putative biometric response in order to determine deviation from the mean. Previous systems provided a stored template and then ranging either side to simply give a pass/fail response. By adaptation of the ranging quotient for a particular individual, a percentage confidence in the pass/fail response can be achieved. For example an individual -may have a wider range of deviation than another individual and therefore close repetition of the biometric data may be indicative of a lower percentage confidence than with an individual with a narrower deviation. Furthermore, through a learning process, the ranging quotient may he varied as a users actual response varies with age or Otherwise.

Claims

1. A method for determining confidence with respect to confirmation of an individual'"'"'s identity by threshold score evaluation of parameters, the method comprising;

(a) separately sampling a parameter given a plurality of times to provide a template;

(b) comparing each sample with the template and a statistical descriptor indicative of a probability range of the difference; and

,

(c) storing a mean similarity score and the statistical descriptor so that upon presentation of a validation parameter set for validation purposes that validation set is compared to the statistical descriptor and the mean similarity score in order to provide a predictive response relative to the likelihood that the parameter is correctly identified.

2. A method as claimed in claim 1 wherein the template consists of a statistical means based on previous sampled parameters.

3. A method as claimed in claim 1 wherein the statistical descriptor is a range quotient indicative of the probability range.

4. A method as claimed in claim 2 wherein the range quotient is a standard deviation limit above the statistical mean.

5. A method as claimed in claim 1 wherein the parameter is a biometric parameter for identification of an individual.

6. A method as claimed in claim 5 wherein a biometric parameter is a signature or handwriting or voice pattern from an individual.

7. A method as claimed in claim 5 wherein each biometric parameter is determined by dynamic stress wave variation as the biometric parameter is formed by the individual.

8. A method as claimed in claim 1 wherein a validation parameter set comprises two or more parameter determinations with respect to response parameters give by an individual for identification.

9. A method as claimed in claim 1 wherein the predictive response is a simple yes or no or pass fail with respect to likelihood that the individual is correctly identified.

10. A method as claimed in claim 1 wherein the predictive response comprises a percentage probability as to the likelihood that the individual is correctly identified.

11. A method as claimed in claim 10 wherein a pre-determined percentage probability value is used to provide a yes/no answer.

12. A method as claimed in claim 1 wherein the validation parameter set in terms of templates is utilised to adapt the mean template and for subsequent determination as to an individual'"'"'s identity.

13. A method as claimed in claim 1 wherein the decision to update a template is based upon the validation parameter.

15. A processor configured to operate in accordance with a method as claimed in claim 1.

14. (canceled)

16. (canceled)