Person recognition apparatus and method thereof

Patent metadata

Field	Value
Publication number	US-9477876-B2
Application number	US-201314026038-A
Country	US
Kind code	B2
Filing date	Sep 13, 2013
Priority date	Sep 20, 2012
Publication date	Oct 25, 2016
Grant date	Oct 25, 2016

Abstract

According to one embodiment, an apparatus includes input unit, detecting unit, extraction unit, storage, selection unit, determination unit, output unit, and setting unit. The selection unit selects operation or setting modes. In operation mode, it is determined whether captured person is preregistered person. In setting mode, threshold for the determination is set. The determination unit determines, as registered person and when operation mode is selected, person with degree of similarity between extracted facial feature information and stored facial feature information of greater than or equal to threshold. The setting unit sets, when setting mode is selected, threshold based on first and second degrees of similarity. First degree of similarity is degree of similarity between facial feature information of the registered person and the stored facial feature information. Second degree of similarity is degree of similarity between facial feature information of person other than registered person and stored facial feature information.

First claim

What is claimed is:

1. A person recognition apparatus comprising:
a facial feature information storage that stores therein facial feature information of a person by each person; and
at least one processor that receives image data captured by a camera;
that detects a face region representing a face of a person from the received image data;
that extracts facial feature information indicating a feature of the face of the person from the detected face region;
that selects one of an operation mode and a setting mode, wherein in the operation mode, it is determined whether the person captured by the camera is a registered person who has been preregistered, and in the setting mode, a threshold value to be used upon the determination is set;
that determines, as the registered person and when the operation mode is selected, a person with a degree of similarity between the extracted facial feature information and the stored facial feature information of greater than or equal to the threshold value;
that outputs a determination result of the at least one processor; and
that sets, when the setting mode is selected, the threshold value based on first degrees of similarity and second degrees of similarity, wherein
the first degrees of similarity are degrees of similarity between facial feature information extracted by capturing the registered person with the camera and the stored facial feature information of the registered person captured,
the second degrees of similarity are degrees of similarity between facial feature information extracted by capturing a person, the facial feature information of the person not previously stored, other than the registered person with the camera and the stored facial feature information,
in the setting mode, the at least one processor selects one of a registrant walking mode and a non-registrant walking mode,
when the registrant walking mode is selected, the at least one processor selects at least a lower one of the first degrees of similarity, and generates a first distribution of similarity based on the first degrees of similarity, and
when the non-registrant walking mode is selected, the at least one processor selects at least a higher one of the second degrees of similarity, and generates a second distribution of similarity based on the second degrees of similarity;
the at least one processor sets, based on the first and second distributions of similarity, the threshold value so that at least one of a rejection rate and an acceptance rate becomes a preset value, the rejection rate being a rate at which the at least one processor does not determine the registered person as a registered person, the acceptance rate is a rate at which the at least one processor determines the person other than the registered person as a registered person.

2. The person recognition apparatus according to claim 1 , wherein,
in the registrant walking mode, the registered person is captured with the camera,
in the non-registrant walking mode, the person other than the registered person is captured with the camera,
when the registrant walking mode is selected, the at least one processor acquires a plurality of the first degrees of similarity to generate the first distribution of similarity,
when the non-registrant walking mode is selected, the at least one processor acquires a plurality of the second degrees of similarity to generate the second distribution of similarity, and.

3. The person recognition apparatus according to claim 2 , wherein, when the registrant walking mode is selected, the at least one processor selects a plurality of frame images having the degrees of similarity which tend to be low from frame images obtained by capturing the registered person with the camera, and generates the first distribution of similarity based on a degree of similarity between facial feature information extracted from the selected frame images and the stored facial feature information.

4. The person recognition apparatus according to claim 2 , wherein, when the non-registrant walking mode is selected, the at least one processor selects a plurality of frame images having the degrees of similarity which tend to be high from frame images obtained by capturing the registered person with the camera, and generates the second distribution of similarity based on a degree of similarity between facial feature information extracted from the selected frame images and the stored facial feature information.

5. The person recognition apparatus according to claim 1 , wherein:
the at least one processor that receives from a user an evaluation indicating at least correctness or incorrectness of the determination result when the at least one processor outputs the determination result while the operation mode is selected; and
the at least one processor that adjusts the threshold value based on the received evaluation.

6. The person recognition apparatus according to claim 5 , further comprising a history storage that stores therein a history of the determination results output by the at least one processor when the operation mode is selected, wherein
the at least one processor receives from the user an evaluation indicating correctness or incorrectness of the stored history of the determination results, and
adjusts the threshold value based on the received evaluation on the stored history of the determination results.

7. A person recognition method performed in a person recognition apparatus that comprises a facial feature information storage that stores therein facial feature information of a person by each person, comprising:
receiving image data captured by a camera;
detecting a face region representing a face of a person from the received image data;
extracting facial feature information indicating a feature of the face of the person from the detected face region;
selecting one of an operation mode and a setting mode, wherein
in the operation mode, it is determined whether the person captured by the camera is a registered person who has been preregistered, and
in the setting mode, a threshold value to be used upon the determination is set;

determining as the registered person and when the operation mode is selected, a person with a degree of similarity between the extracted facial feature information and the stored facial
feature information is greater than or equal to the threshold value;
outputting a determination result; and
setting when the setting mode is selected, the threshold value based on first degrees of similarity and second degrees of similarity, wherein
the first degrees of similarity are degrees of similarity between facial feature information extracted by capturing the registered person with the camera and the stored facial feature information of the registered person captured, and
the second degrees of similarity are degrees of similarity between facial feature information extracted by capturing a person, the facial feature information of the person not previously stored, other than the registered person with the camera and the stored facial feature information, wherein

in the setting mode, the at least one processor selects one of a registrant walking mode and a non-registrant walking mode,
when the registrant walking mode is selected, the at least one processor selects at least a lower one of the acquired first degrees of similarity, and generates a first distribution of similarity based on the first degrees of similarity, and
when the non-registrant walking mode is selected, the at least one processor selects at least a higher one of the acquired second degrees of similarity, and generates a second distribution of similarity based on the second degrees of similarity,
the at least one processor sets, based on the first and second distrib

Assignees

• Toshiba Kk

Inventors

• Sukegawa Hiroshi
• Saito Hiroo

Classifications

• G06K9/00288 Primary

  Physics · mapped topic

• G06V40/172 Primary

  Classification, e.g. identification · CPC title