Multiple-input, automatic recognition method and apparatus

• US 7,664,962 B2
• Filed: 03/13/2006
• Issued: 02/16/2010
• Est. Priority Date: 03/13/2006
• Status: Expired due to Fees

• Current Assignee: Motorola Solutions, Inc.
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Kuhlman, Douglas
• Primary Examiner(s): Simitoski; Michael J
• Assistant Examiner(s): Durham; Imhotep
• Application Number:
  US11/276,732
• Publication Number:
  US 20070210895A1
• Time in Patent Office: 1,436 Days
• Field of Search: 382115-127, 713/186, 726/22, 706/15
• US Class Current: 713/186
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition

Abstract

Automatic recognition apparatus (100, 700) includes multiple automatic recognition subsystems (102, 104, 106) that are cable of producing estimates of the probability that a subject matches a particular identity, a probability estimate combiner (108) that receives estimates from the multiple automatic recognition subsystems (102, 104, 106) and produces a combined estimate, and a decayer that (111) that decreases the certainty of the combined estimate over time while avoiding changing estimates to values that would suggest strong information contrary to the initial values of the combined estimates.

Claims

1. A computer-readable medium storing programming instructions for decaying an estimate of a probability that a subject has a predetermined identity, the computer-readable medium including programming instructions for:

calculating a series of digit values from the estimate of the probability by iteratively applying a halving function to the estimate of the probability;

multiplying each digit value by a corresponding digit of an output of a time-dependent decay function to obtain a series of terms;

iteratively combining the series of terms with a combiner function that accepts two inputs and produces an output, wherein an inverse of the combiner function when applied to combine a variable with itself is the halving function, whereby a decayed estimate of the probability that the subject has the predetermined identity is calculated; and

selectively performing, based on the decayed estimate, a function selected from the group consisting of;

granting access to a secure resource and granting access to a secure facility.

2. The computer-readable medium according to claim 1 wherein said decay function is an exponential function.

3. The computer-readable medium according to claim 1:

wherein the combiner function is selected from the group consisting of;

4. A computer-readable medium storing programming instructions for decaying an estimate of a probability that a subject has a predetermined identity, the computer-readable medium including programming instructions for:

calculating a series of digit values from the estimate of the probability by iteratively applying a halving function to the estimate of the probability;

multiplying each digit value by a corresponding digit of an output of a time-dependent decay function to obtain a series of terms;

iteratively combining the series of terms with a combiner function that accepts two inputs and produces an output, whereby a decayed estimate of the probability that the subject has the predetermined identity is calculated; and

selectively performing, based on the decayed estimate, a function selected from the group consisting of;

granting access to a secure resource and granting access to a secure facility;

wherein the combining function and the halving function are used on a predetermined domain that includes a predetermined interior point that represents an equal probability that the subject does and does not match the predetermined identity;

wherein the predetermined interior point divides the predetermined domain into a first subdomain in which it is more likely that the subject does not have the predetermined identity and a second subdomain in which it is more likely that the subject does have the predetermined identity;

wherein, for input values in the first subdomain, the halving function returns function values that are in the first domain and are closer to the predetermined interior point than the input values in the first subdomain; and

wherein, for input values in the second subdomain, the halving function returns function values that are in the second domain and are closer to the predetermined interior point than the input values in the second subdomain.

5. The computer-readable medium according to claim 4 wherein said decay function is an exponential function.

6. The computer-readable medium according to claim 4:

wherein the combiner function is selected from the group consisting of;

7. A system for decaying an estimate of a probability that a subject has a predetermined identity, the system comprising:

a decayer, comprising a processor unit, adapted to;

calculate a series of digit values from the estimate of the probability by iteratively applying a halving function to the estimate of the probability;

multiply each digit value by a corresponding digit of an output of a time-dependent decay function to obtain a series of terms; and

iteratively combine the series of terms with a combiner function that accepts two inputs and produces an output, wherein an inverse of the combiner function when applied to combine a variable with itself is the halving function, whereby a decayed estimate of the probability that the subject has the predetermined identity is calculated; and

a subsystem adapted to;

selectively perform, based on the decayed estimate, a function selected from the group consisting of;

granting access to a secure resource and granting access to a secure facility.

8. The system according to claim 7 wherein said decay function is an exponential function.

9. The system according to claim 7:

wherein the combiner function is selected from the group consisting of;

10. A system for decaying an estimate of a probability that a subject has a predetermined identity, the system comprising:

a decayer, comprising a processor unit, adapted to;

calculate a series of digit values from the estimate of the probability by iteratively applying a halving function to the estimate of the probability;

multiply each digit value by a corresponding digit of an output of a time-dependent decay function to obtain a series of terms; and

iteratively combine the series of terms with a combiner function that accepts two inputs and produces an output, whereby a decayed estimate of the probability that the subject has the predetermined identity is calculated; and

a subsystem adapted to;

selectively perform, based on the decayed estimate, a function selected from the group consisting of;

granting access to a secure resource and granting access to a secure facility;

wherein the combining function and the halving function are used on a predetermined domain that includes a predetermined interior point that represents an equal probability that the subject does and does not match the predetermined identity;

wherein the predetermined interior point divides the predetermined domain into a first subdomain in which it is more likely that the subject does not have the predetermined identity and a second subdomain in which it is more likely that the subject does have the predetermined identity;

wherein, for input values in the first subdomain, the halving function returns function values that are in the first domain and are closer to the predetermined interior point than the input values in the first subdomain; and

wherein, for input values in the second subdomain, the halving function returns function values that are in the second domain and are closer to the predetermined interior point than the input values in the second subdomain.

11. The system according to claim 10 wherein said decay function is an exponential function.

12. The system according to claim 10:

wherein the combiner function is selected from the group consisting of;