Image processing device, image processing method, and storage medium storing image processing program

• US 9,262,614 B2
• Filed: 08/01/2013
• Issued: 02/16/2016
• Est. Priority Date: 10/19/2012
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Yamaguchi, Nobuyasu
• Primary Examiner(s): Koziol, Stephen R
• Assistant Examiner(s): Gilliard, Delomia L
• Application Number:
  US13/956,738
• Publication Number:
  US 20140112553A1
• Time in Patent Office: 929 Days
• Field of Search: 382/118
• US Class Current: 1/1
• CPC Class Codes: G06F 21/32 using biometric data, e.g. fingerprints, iris scans or voiceprints
• CPC Class Codes: G06F 2221/2139 Recurrent verification
• 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 40/162 using pixel segmentation or colour matching
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

An image processing device includes a processor; and a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute, acquiring a captured image; extracting a first feature amount and a second feature amount of a user included in the image, and a third feature amount indicative of a likelihood of a portion from which the second feature amount is extracted; generating an authentication table including a plurality of the first feature amounts, a plurality of the second feature amounts, and a plurality of the third feature amounts extracted from a plurality of the images; and selecting the third feature amount extracted at authentication of the user, and the first feature amount and the second feature amount included in the authentication table, based on the third feature amount included in the authentication table, and authenticating the user based on the first feature amount.

Claims

1. An image processing device comprising:

a processor; and

a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute,

acquiring a plurality of captured images;

extracting, from each of the plurality of captured images, a first feature amount and a second feature amount of a user included in the captured image, and a third feature amount indicating a condition of acquiring of the captured image;

generating an authentication table including a plurality of templates, each of which corresponds to each of the plurality of the captured images, each of the templates including the first feature amount, the second feature amount, and the third feature amount extracted from each of the plurality of captured images;

acquiring an authentication image of a target user;

extracting the third feature amount from the authentication image of the target user;

selecting a template among the plurality of the templates by comparing the third feature amount of the authentication image and a plurality of third feature amounts included in the authentication table; and

authenticating the target user based on the first feature amount and the second feature amount which are included in the selected template, and the first feature amount and the second feature amount which are extracted from the authentication image.

2. The device according to claim 1,

wherein the first feature amount is a color histogram, and the second feature amount is a feature amount vector.

3. The device according to claim 1,

wherein the generating classifies the authentication table into a first group set including a plurality of first groups based on a degree of the third feature amount and a predetermined first threshold, and

wherein the authenticating selects a second group from a second group set belonging to one of the first group set and including a plurality of the second groups based on the third feature amount extracted at the authentication, and authenticates the target user based on the first feature amount and the second feature amount belonging to the selected second group, and the first feature amount and the second feature amount extracted at the authentication.

4. The device according to claim 3,

wherein the generating classifies the second group set into the plurality of second groups each belonging to the first group based on a first similarity between the plurality of first feature amounts and a predetermined second threshold, and a second similarity between the plurality of second feature amounts and a predetermined third threshold.

5. The device according to claim 4,

wherein the generating classifies one of the second groups based on the first similarity with respect to the first feature amount with the first similarity being maximum, and the second threshold, and classifies another one of the second groups based on the first similarity with respect to the first feature amount with the first similarity being maximum from among the plurality of first feature amounts not classified into the one second group, and the second threshold.

6. The device according to claim 4,

wherein the generating classifies one of the second groups based on the second similarity with respect to the second feature amount with the second similarity being maximum, and the third threshold, and classifies another one of the second groups based on the second similarity with respect to the second feature amount with the second similarity being maximum from among the plurality of second feature amounts not classified into the one second group, and the third threshold.

7. The device according to claim 3,

wherein the authenticating authenticates the target user by using the first feature amount with an average or a sum total of the first similarity being maximum, or the second feature amount with an average or a sum total of the second similarity being maximum in the second group.

8. The device according to claim 1,

wherein the authenticating excludes the selected first feature amount or the selected second feature amount from a selection subject of the authentication table when the target user is not authenticated, and repetitively performs the authentication based on a false acceptance rate determined as a requirement of the image processing device and a false acceptance rate per single authentication processing using the second feature amount.

9. The device according to claim 1,

wherein the authenticating excludes the selected first feature amount or the selected second feature amount from a selection subject of the authentication table when the target user is not authenticated, and repetitively performs the authentication based on a false acceptance rate determined as a requirement of the image processing device and a false acceptance rate per single authentication processing using the first feature amount.

20. The device according to claim 1, wherein the third feature amount indicates an orientation of the user in the captured image.

10. An image processing method comprising:

acquiring a plurality of captured images;

extracting, from each of the plurality of captured images, a first feature amount and a second feature amount of a user included in the captured image, and a third feature amount indicating a condition of acquiring of the captured image;

generating, by a computer processor, an authentication table including a plurality of templates, each of which corresponds to each of the plurality of the captured images, each of the templates including the first feature amount, the second feature amount, and the third feature amount extracted from each of the plurality of captured images;

acquiring an authentication image of a target user;

extracting the third feature amount from the authentication image of the target user;

selecting a template among the plurality of the templates by comparing the third feature amount of the authentication image and a plurality of third feature amounts included in the authentication table; and

authenticating the target user based on the first feature amount and the second feature amount which are included in the selected template, and the first feature amount and the second feature amount which are extracted from the authentication image.

11. The method according to claim 10,

wherein the first feature amount is a color histogram, and the second feature amount is a feature amount vector.

12. The method according to claim 10,

wherein the generating classifies the authentication table into a first group set including a plurality of first groups based on a degree of the third feature amount and a predetermined first threshold, and

wherein the authenticating selects a second group from a second group set belonging to one of the first group set and including a plurality of the second groups based on the third feature amount extracted at the authentication, and authenticates the target user based on the first feature amount and the second feature amount belonging to the selected second group, and the first feature amount and the second feature amount extracted at the authentication.

13. The method according to claim 12,

wherein the generating classifies the second group set into the plurality of second groups each belonging to the first group based on a first similarity between the plurality of first feature amounts and a predetermined second threshold, and a second similarity between the plurality of second feature amounts and a predetermined third threshold.

14. The method according to claim 13,

wherein the generating classifies one of the second groups based on the first similarity with respect to the first feature amount with the first similarity being maximum, and the second threshold, and classifies another one of the second groups based on the first similarity with respect to the first feature amount with the first similarity being maximum from among the plurality of first feature amounts not classified into the one second group, and the second threshold.

15. The method according to claim 13,

wherein the generating classifies one of the second groups based on the second similarity with respect to the second feature amount with the second similarity being maximum, and the third threshold, and classifies another one of the second groups based on the second similarity with respect to the second feature amount with the second similarity being maximum from among the plurality of second feature amounts not classified into the one second group, and the third threshold.

16. The method according to claim 12,

wherein the authenticating authenticates the user by using the first feature amount with an average or a sum total of the first similarity being maximum, or the second feature amount with an average or a sum total of the second similarity being maximum in the second group.

17. The method according to claim 10,

wherein the authenticating excludes the selected first feature amount or the selected second feature amount from a selection subject of the authentication table when the target user is not authenticated, and repetitively performs the authentication based on a false acceptance rate determined as a requirement of the image processing device and a false acceptance rate per single authentication processing using the second feature amount.

18. The method according to claim 10,

wherein the authenticating excludes the selected first feature amount or the selected second feature amount from a selection subject of the authentication table when the target user is not authenticated, and repetitively performs the authentication based on a false acceptance rate determined as a requirement of the image processing device and a false acceptance rate per single authentication processing using the first feature amount.

19. A non-transitory computer-readable storage medium storing an image processing program that causing a computer to execute a process comprising:

acquiring a plurality of captured images;

extracting, from each of the plurality of captured images, a first feature amount and a second feature amount of a user included in the captured image, and a third feature amount indicating a condition of acquiring of the captured image;

generating an authentication table including a plurality of templates, each of which corresponds to each of the plurality of the captured images, each of the templates including the first feature amount, the second feature amount, and the third feature amount extracted from each of the plurality of captured images;

acquiring an authentication image of a target user;

extracting the third feature amount from the authentication image of the target user;

selecting a template among the plurality of the templates by comparing the third feature amount of the authentication image and a plurality of third feature amounts included in the authentication table; and

authenticating the target user based on the first feature amount and the second feature amount which are included in the selected template, and the first feature amount and the second feature amount which are extracted from the authentication image.