Biometric identification system using pulse waveform

• US 9,223,298 B2
• Filed: 06/13/2013
• Issued: 12/29/2015
• Est. Priority Date: 04/04/2011
• Status: Active Grant

• Current Assignee: Lifeloc Technologies Inc. (Edco Partners LLLP)
• Original Assignee: Integrated Monitoring Systems, LLC (Track Group, Inc.)
• Inventors: Phillips, Brian K., Wilson, Geoffrey A.
• Primary Examiner(s): Zimmerman, Brian
• Assistant Examiner(s): Patel, Bhavin M
• Application Number:
  US13/916,818
• Publication Number:
  US 20130271263A1
• Time in Patent Office: 929 Days
• Field of Search: 340/5.82
• US Class Current: 1/1
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/026 Measuring blood flow A61B3/1233, A61B3/1241 take precedence
• CPC Class Codes: A61B 5/117 Identification of persons methods or arrangements for recognising patterns, e.g. fingerprints, G06F18/00, G06V40/00; identification of persons by analysing their voice or speech G10L17/00
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: G05B 1/00 Comparing elements, i.e. elements for effecting comparison directly or indirectly between a desired value and existing or anticipated values

Abstract

A biometric identity confirmation system is based on pulse waveform data for the subject. During an initial enrollment mode, pulse waveform data for a known subject are used to generate subject characterization data for the known subject. The subject characterization data includes an exemplar created by synchronous averaging of pulse waveform data over multiple pulse cycles. A number of trigger candidate are identified for the start point of each pulse cycle. The time delay between trigger candidates is analyzed to discard false trigger candidate and identify true candidates, which are then used as the start points for each pulse cycle in synchronous averaging of the pulse waveform data. During a subsequent identity authentication mode, pulse waveform data for a test subject are analyzed using the subject characterization data to confirm whether the identity of the test subject matches the known subject.

Claims

1. A method for biometric identity confirmation of a subject having a pulse, said method comprising:

during an initial training mode, acquiring pulse waveform data from a known subject;

generating and storing subject characterization data for the known subject based at least in part on an exemplar created by synchronous averaging of pulse waveform data over multiple pulse cycles, wherein said synchronous averaging of pulse waveform data includes;

(a) identifying trigger candidates for a start point of each pulse cycle;

(b) analyzing a time delay between trigger candidates to discard false trigger candidates and identify true trigger candidates; and

(c) synchronously averaging the pulse waveform data for each pulse cycle using the true trigger candidates as start points for each pulse cycle; and

during a subsequent identity authentication mode, acquiring pulse waveform data from a test subject, and analyzing the pulse waveform data with the subject characterization data for the known subject to confirm whether the identity of the test subject matches the known subject.

2. The method of claim 1 wherein the trigger candidates are derived at least in part from a second derivative of the pulse waveform data with respect to time.

3. The method of claim 1 wherein the step of acquiring and analyzing pulse waveform data for a test subject further comprises:

acquiring pulse waveform data from the test subject over multiple pulse cycles;

identifying trigger candidates for the start point of each pulse cycle;

analyzing the time delay between trigger candidates to discard false trigger candidates and identify true trigger candidates; and

synchronously averaging the pulse waveform data for each pulse cycle using the true trigger candidates as the start points for each pulse cycle.

4. The method of claim 1 wherein the step of generating subject characterization data further comprises:

computing an exemplar in the form of a parameter vector from the pulse waveform data for the known subject;

computing a covariance matrix from the pulse wave data for the known subject;

computing an optimal composite parameter from the covariance matrix and parameter vector that is characteristic of the known subject; and

computing an authentication threshold corresponding to a desired true authentication probability for the known subject.

5. The method of claim 1 wherein the step of analyzing the pulse waveform data with the subject characterization data for the known subject to confirm whether the identity of the test subject matches the known subject further comprises:

computing a deviation of the pulse waveform data for the test subject from the exemplar for the known subject;

computing an optimal composite parameter from the deviation; and

confirming the identity of the test subject matches the known subject if optimal composite parameter is greater than the authentication threshold for the known subject.

6. The method of claim 1 wherein the step of generating subject characterization data further comprises:

computing an exemplar in the form of a parameter vector from the pulse waveform data for the known subject;

computing a covariance matrix from the pulse waveform data for the known subject; and

computing a probability distribution ratio of a weighted subject/impostor probability density by a Bayesian optimal decision analysis of the parameter vector, covariance matrix, and data from other subjects as potential impostors for the known subject.

7. The method of claim 1 wherein the step of analyzing the pulse waveform data with the subject characterization data for the known subject to confirm whether the identity of the test subject matches the known subject further comprises:

computing a deviation of the pulse waveform data for the test subject from the exemplar for the known subject;

computing a weighted subject/impostor probability density ratio for the deviation; and

confirming the identity of the test subject matches the known subject if the weighted subject/impostor probability density ratio is greater than one.

8. The method of claim 1 wherein the step of generating subject characterization data further comprises weighting portions of the exemplar selected based on their repeatability observed during the initial training mode.

9. The method of claim 1 wherein the step of generating subject characterization data further comprises weighting portions of the exemplar selected to distinguish characteristic features of the known subject observed during the initial training mode.

10. The method of claim 1 wherein the initial training mode acquires a plurality of sets of pulse waveform data over a plurality of trials, and wherein the step of synchronous averaging includes only the pulse waveform data from the most consistent trials.

11. The method of claim 1 wherein the initial training mode acquires a plurality of sets of pulse waveform data over a plurality of trials, and wherein the trials are rejected if the variance of the pulse waveform data exceeds a predetermined threshold.