Personal identification device, personal identification method, updating method for identification dictionary data, and updating program for identification dictionary data

• US 7,916,903 B2
• Filed: 11/09/2007
• Issued: 03/29/2011
• Est. Priority Date: 11/21/2006
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Eura, Yuka, Kojima, Tamaki
• Primary Examiner(s): Tabatabai, Abolfazl
• Application Number:
  US11/983,691
• Publication Number:
  US 20080123907A1
• Time in Patent Office: 1,236 Days
• Field of Search: 382/100, 382/115, 382/116, 382/117, 382/118, 348/154, 348/155, 348/169, 348/170, 348/171, 348/172, 348/267, 379/93.03, 379/207.13
• US Class Current: 382/118
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A personal identification device compares personal identification dictionary data, which is generated from feature data extracted from an identification element for identifying a person and is stored in association with person-specifying information for specifying a person, against feature data extracted from an identification element extracted from a person to thereby identify a person having the captured identification element. The personal identification device includes a dictionary data update processing section that assigns a weighting factor, which is generated from capture time information newly acquired by a time acquiring section and capture time information of an identification element acquired previously and stored in a dictionary-data generating information storing section in association with person-specifying information, to feature data captured at each capture time to thereby generate or update identification dictionary data corresponding to person-specifying information acquired by a person-specifying section on the basis of the result of personal identification.

Claims

1. A personal identification device comprising:

a dictionary data storing section that stores identification dictionary data for a person to be identified, which is generated from feature data extracted from an identification element for identifying the person, in association with person-specifying information for specifying the person;

identification means for comparing feature data extracted from an identification element extracted from the person, against the identification dictionary data in the dictionary data storing section to identify a person who has the captured identification element;

time acquiring means for acquiring information on capture time of an identification element captured from the person;

person-specifying means for obtaining the person-specifying information that specifies a person who has the captured identification element, on the basis of a result of identification of the person by the identification means;

dictionary-data generating information storing means for adding and storing data of the captured identification element or feature data extracted from the identification element, and time information acquired by the time acquiring means into the dictionary-data generating information storing section, in association with the person-specifying information acquired by the person-specifying means; and

dictionary data update processing means for generating or updating the identification dictionary data corresponding to the person-specifying information acquired by the person-specifying means, by using information on capture time newly acquired by the time acquiring means, information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying means, and feature data captured at each capture time, the dictionary data update processing means being configured to update the identification dictionary data by weighting each piece of the feature data by a weighting factor according to its capture time.

2. The personal identification device according to claim 1, wherein the dictionary data update processing means includes:

weighting factor computing means for computing a weighting factor for feature data with respect to the newly captured identification element, and a weighting factor for feature data with respect to the identification element captured previously, by using time information newly acquired by the time acquiring means and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying means; and

dictionary data generating or updating means for assigning each of the weighting factors computed by the weighting factor computing means to feature data of the identification element captured previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying means, and to feature data of the newly acquired identification element, and for generating or updating the identification dictionary data stored in association with the person-specifying information acquired by the person-specifying means, by using the plurality of pieces of weighted feature data.

3. The personal identification device according to claim 1, wherein the dictionary data update processing means includes:

weighting factor computing means for computing a weighting factor for feature data with respect to the newly captured identification element of a person, by using time information newly acquired by the time acquiring means and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying means, and for adding and storing the computed weighting factor into the dictionary-data generating information storing section; and

dictionary data generating or updating means for assigning a corresponding weighting factor stored in the dictionary-data generating information storing section to feature data of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying means, assigning a weighting factor computed by the weighting factor computing means to feature data of the newly acquired identification element, and generating or updating the identification dictionary data stored in association with the person-specifying information acquired by the person-specifying means, by using the plurality of pieces of weighted feature data.

4. The personal identification device according to claim 1, wherein:

the dictionary-data generating information storing section can store a finite number of pieces of data of an identification element of the person which can be stored in association with the person-specifying information or a finite number of pieces of feature data extracted from the identification element, and a finite number of pieces of time information acquired by the time acquiring means; and

after storing the finite number of pieces of data or feature data of the identification element which can be stored in the dictionary-date generating information storing section in association with the person-specifying information, and the acquired time information, when an identification element of the person is newly captured, data with the earliest capture time is discarded, and data or feature data of the newly captured identification element and the acquired time information are stored into the dictionary-data generating information storing section.

5. The personal identification device according to claim 1, wherein:

the weighting factor computing means computes the weighting factor with respect to each piece of the feature data on the basis of capture time of the identification element such that the weighting factor increases linearly toward the latest point in a time series.

6. The personal identification device according to claim 1, wherein:

the weighting factor computing means computes the weighting factor with respect to each piece of the feature data on the basis of capture time of the identification element such that the weighting factor increases non-linearly toward the latest time point in a time series.

7. The personal identification device according to claim 1, wherein:

the weighting factor computing means computes the weighting factor with respect to each piece of the feature data on the basis of a time difference between each capture time of the identification element and the earliest capture time of the identification element.

8. The personal identification device according to claim 1, wherein:

the weighting factor computing means divides a period of time including the earliest and latest capture times of the identification element into a plurality of predetermined time length sections, counts the number of capture times of the identification element included in each of the predetermined time length sections, and computes the weighting factor with respect to each piece of the feature data for each of the predetermined time length sections such that the weighting factor increases as the count value becomes larger.

9. The personal identification device according to claim 1, wherein:

when the identification element is newly captured, the weighting factor computing means computes a weighting factor with respect to feature data of the newly captured identification element on the basis of a time difference between capture time of the newly captured identification element, and capture time of the identification element captured immediately previous to the newly captured identification element.

10. The personal identification device according to claim 1, wherein:

the personal identification device includes a plurality of weighting factor computing methods for use by the weighting factor computing means; and

the personal identification device further comprises means for inquiring a user which of the plurality of weighting factor computing methods is to be used, and accepting a user'"'"'s selection and setting with respect to the inquiry,

wherein the weighting factor computing means computes the weighting factor by using the weighting factor computing method selected and set by the user.

11. A personal identification method comprising:

an identification step of comparing feature data extracted from an identification element extracted from a person, against identification dictionary data of the person to be identified which is generated from the feature data to identify a person who has the captured identification element, the identification dictionary data being stored in a dictionary data storing section in association with person-specifying information for specifying the person;

a time acquiring step of acquiring information on capture time of an identification element captured from the person;

a person-specifying step of obtaining the person-specifying information that specifies a person who has the captured identification element, on the basis of a result of identification of the person in the identification step;

a dictionary-data generating information storing step of adding and storing data of the captured identification element or feature data extracted from the identification element, and time information acquired in the time acquiring step into the dictionary-data generating information storing section, in association with the person-specifying information acquired in the person-specifying step; and

a dictionary data update processing step of generating or updating the identification dictionary data corresponding to the person-specifying information acquired in the person-specifying means, by using information on capture time newly acquired in the time acquiring step, information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and feature data captured at each capture time, the dictionary data update processing step being configured to update the identification dictionary data by weighting each piece of the feature data by a weighting factor according to its capture time.

12. The personal identification method according to claim 11, wherein the dictionary data update processing step includes:

a weighting factor computing step of computing a weighting factor for feature data with respect to the newly captured identification element, and a weighting factor for feature data with respect to the identification element captured previously, by using time information newly acquired in the time acquiring step and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step; and

a dictionary data generating or updating step of assigning each of the weighting factors computed in the weighting factor computing step to feature data of the identification element captured previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and to feature data of the newly acquired identification element, and generating or updating the identification dictionary data stored in association with the person-specifying information acquired in the person-specifying step, by using the plurality of pieces of weighted feature data.

13. The personal identification method according to claim 11, wherein the dictionary data update processing step includes:

a weighting factor computing step of computing a weighting factor for feature data with respect to the newly captured identification element of a person, by using time information newly acquired in the time acquiring step and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and for adding and storing the computed weighting factor into the dictionary-data generating information storing section; and

a dictionary data generating or updating step of assigning a corresponding weighting factor stored in the dictionary-data generating information storing section to feature data of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, assigning a weighting factor computed by the weighting factor computing step to feature data of the newly acquired identification element, and generating or updating the identification dictionary data stored in association with the person-specifying information acquired in the person-specifying step, by using the plurality of pieces of weighted feature data.

14. An updating method for identification dictionary data in a personal identification device that compares identification dictionary data of a person to be identified, which is generated from feature data extracted from an identification element for identifying the person and stored in association with person-specifying information for specifying the person, against feature data extracted from an identification element captured from a person to identify a person who has the captured identification element, comprising:

a time acquiring step of acquiring information on capture time of an identification element captured from the person;

a person-specifying step of obtaining the person-specifying information that specifies a person who has the captured identification element, on the basis of a result of identification of a person who has the captured identification element;

a dictionary-data generating information storing step of adding and storing data of the captured identification element of a person or feature data extracted from the identification element, and time information acquired in the time acquiring step into the dictionary-data generating information storing section, in association with the person-specifying information acquired in the person-specifying step; and

a dictionary data update processing step of generating or updating the identification dictionary data corresponding to the person-specifying information acquired in the person-specifying means, by using information on capture time newly acquired in the time acquiring step, information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and feature data captured at each capture time, the dictionary data update processing step being configured to update the identification dictionary data by weighting each piece of the feature data by a weighting factor according to its capture time.

15. The updating method for identification dictionary data according to claim 14, wherein the dictionary data update processing step includes:

a weighting factor computing step of computing a weighting factor for feature data with respect to the newly captured identification element, and a weighting factor for feature data with respect to the identification element captured previously, by using time information newly acquired in the time acquiring step and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step; and

a dictionary data generating or updating step of assigning each of the weighting factors computed in the weighting factor computing step to feature data of the identification element captured previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and to feature data of the newly acquired identification element, and generating or updating the identification dictionary data stored in association with the person-specifying information acquired in the person-specifying step, by using the plurality of pieces of weighted feature data.

16. The updating method for identification dictionary data according to claim 14, wherein the dictionary data update processing step includes:

a weighting factor computing step of computing a weighting factor for feature data with respect to the newly captured identification element of a person, by using time information newly acquired in the time acquiring step and information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and for adding and storing the computed weighting factor into the dictionary-data generating information storing section; and

a dictionary data generating or updating step of assigning a corresponding weighting factor stored in the dictionary-data generating information storing section to feature data of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, assigning a weighting factor computed by the weighting factor computing step to feature data of the newly acquired identification element, and generating or updating the identification dictionary data stored in association with the person-specifying information acquired in the person-specifying step, by using the plurality of pieces of weighted feature data.

17. An updating program stored on a computer readable medium for identification dictionary data in a personal identification device that compares identification dictionary data of a person to be identified, which is generated from feature data extracted from an identification element for identifying the person and stored in association with person-specifying information for specifying the person, against feature data extracted from a captured identification element of a person to identify a person who has the captured identification element, the updating program being configured to generate or update the identification dictionary data by causing a computer to execute:

a time acquiring step of acquiring information on capture time of an identification element captured from the person;

a person-specifying step of obtaining the person-specifying information that specifies a person who has the captured identification element, on the basis of a result of identification of a person who has the captured identification element;

a dictionary-data generating information storing step of adding and storing data of the captured identification element of a person or feature data extracted from the identification element, and time information acquired in the time acquiring step into the dictionary-data generating information storing section, in association with the person-specifying information acquired in the person-specifying step; and

a dictionary data update processing step of generating or updating the identification dictionary data corresponding to the person-specifying information acquired in the person-specifying means, by using information on capture time newly acquired in the time acquiring step, information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired in the person-specifying step, and feature data captured at each capture time, the dictionary data update processing step being configured to update the identification dictionary data by weighting each piece of the feature data by a weighting factor according to its capture time.

18. A personal identification device comprising:

a processor to perform the following;

a dictionary data storing section that stores identification dictionary data for a person to be identified, which is generated from feature data extracted from an identification element for identifying the person, in association with person-specifying information for specifying the person;

an identification section that compares feature data extracted from an identification element extracted from the person, against the identification dictionary data in the dictionary data storing section to identify a person who has the captured identification element;

a time acquiring section that acquires information on capture time of an identification element captured from the person;

a person-specifying section that obtains the person-specifying information that specifies a person who has the captured identification element, on the basis of a result of identification of the person by the identification section;

a dictionary-data generating information storing section that adds and stores data of the captured identification element or feature data extracted from the identification element, and time information acquired by the time acquiring section into the dictionary-data generating information storing section, in association with the person-specifying information acquired by the person-specifying section; and

a dictionary data update processing section that generates or updates the identification dictionary data corresponding to the person-specifying information acquired by the person-specifying section, by using information on capture time newly acquired by the time acquiring section, information on capture time of the identification element acquired previously and stored in the dictionary-data generating information storing section in association with the person-specifying information acquired by the person-specifying section, and feature data captured at each capture time, the dictionary data update processing section being configured to update the identification dictionary data by weighting each piece of the feature data by a weighting factor according to its capture time.