Reliability acquiring apparatus, reliability acquiring method, and reliability acquiring program

• US 9,858,501 B2
• Filed: 02/08/2013
• Issued: 01/02/2018
• Est. Priority Date: 02/16/2012
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Morishita, Yusuke
• Primary Examiner(s): Rahmjoo, Manuchehr
• Application Number:
  US14/379,304
• Publication Number:
  US 20150023606A1
• Time in Patent Office: 1,789 Days
• Field of Search: 382219
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06V 40/166 using acquisition arrangements
• CPC Class Codes: G06V 40/167 using comparisons between temporally consecutive images
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships

Abstract

A reliability acquiring apparatus includes a section that stores information on a classifier that outputs part likelihood concerning a predetermined part of a detection target when applied to an image, a section that calculates, concerning an image region included in an input image, the part likelihood on the basis of the information on the classifier, a section that determines, on the basis of the calculated part likelihood, a position of the predetermined part in the input image, a section that stores information on a reference position of the predetermined part, a section that calculates, on the basis of the information on the reference position, difference information between the reference position of the predetermined part and the determined position of the predetermined part, and a section that calculates, on the basis of the difference information, reliability indicating possibility that the input image is an image of the detection target.

Claims

1. A reliability calculating apparatus for detection of a target as a face, comprising:

a processor; and

a storing device,

wherein the storing device comprises;

classifier storage that stores information on a classifier that outputs a part likelihood of a predetermined part of a detection target which is calculated based on image feature values concerning the predetermined part of the detection target, and

reference position storage that stores information on a reference position of the predetermined part,

wherein the storing device further has, stored therein, programming code that, upon execution by the processor, causes the processor to perform the functions of;

retrieving, from the classifier storing device, information on the classifier from the classifier storage,

calculating, using data of an image region included in an input image, a part likelihood of the input image by using the retrieved information on the classifier,

determining, based on the calculated part likelihood, a position of the predetermined part in the input image,

retrieving information on the reference position of the predetermined part of the detection target stored from the reference position information storing,

calculating difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part, and

calculating first reliability indicating possibility that the input image is an image showing the detection target based on a facial feature point likelihood calculated based on the image feature values,

wherein the image feature values include a difference in intensity values,

wherein the programming code causes the processor, in calculating difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part, to calculate the difference information on the basis of whether the position of the predetermined part in the input image determined by the position determining unit is an outlier using a method of robust estimation,

and wherein the detection target is a face.

2. The reliability calculating apparatus according to claim 1,

wherein the programming code further causes the processor to calculate a third reliability based on the first reliability and one or more second reliabilities, and

wherein each of the second reliabilities indicates a possibility that the input image is the image of the detection target, and is calculated in a manner different from that of calculating the first reliability.

3. The reliability calculating apparatus according to claim 1, wherein

the classifier storage and the reference position information storage respectively store information on part likelihood and information on reference positions for a plurality of predetermined parts, and

the part likelihood acquiring unit, the part position determining unit, and the difference acquiring unit, respectively concerning the plurality of predetermined parts, acquire the part likelihood, determine positions of the predetermined parts, and calculate the difference information.

4. The reliability calculating apparatus according to claim 3, wherein the programming code, in calculating the first reliability, causes the processor to calculate the reliability on the basis of the difference information and the calculated part likelihood.

5. The reliability calculating apparatus according to claim 4,

wherein the programming code further causes the processor to calculate a third reliability based on the first reliability and one or more second reliabilities, and

wherein each of the second reliabilities indicates a possibility that the input image is the image of the detection target, and is calculated in a manner different from that of calculating the first reliability.

6. The reliability calculating apparatus according to claim 3,

wherein the programming code further causes the processor to calculate a third reliability based on the first reliability and one or more second reliabilities, and

wherein each of the second reliabilities indicates a possibility that the input image is the image of the detection target, and is calculated in a manner different from that of calculating the first reliability.

7. The reliability calculating apparatus according to claim 1, wherein the programming code, in calculating the first reliability, causes the processor to calculate the reliability on the basis of the difference information and the calculated part likelihood.

8. The reliability calculating apparatus according to claim 7,

wherein the programming code further causes the processor to calculate a third reliability based on the first reliability and one or more second reliabilities, and

wherein each of the second reliabilities indicates a possibility that the input image is the image of the detection target, and is calculated in a manner different from that of calculating the first reliability.

9. The reliability calculating apparatus according to claim 1,

wherein the programming code further causes the processor to calculate a third reliability based on the first reliability and one or more second reliabilities, and

wherein each of the second reliabilities indicates a possibility that the input image is the image of the detection target, and is calculated in a manner different from that of calculating the first reliability.

10. A reliability calculating method for detection of a target as a face performed by a computer, comprising the computer-performed steps of:

retrieving, from a classifier storing device, information on a classifier which outputs a part likelihood of a predetermined part of a detection target calculated based on image feature values concerning the predetermined part of the detection target;

calculating, using data of an image region included in an input image, a part likelihood of the input image by using the retrieved information on the classifier;

determining a position, in the input image, of the predetermined part of the detection target based on the calculated part likelihood;

retrieving information on a reference position of the predetermined part of the detection target stored in a reference position information storing device;

calculating difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part; and

calculating a first reliability that indicates a possibility that the input image is an image showing the detection target based on a facial feature point likelihood calculated based on the image feature values,

wherein the image feature values include a difference in intensity values,

wherein, in calculating the difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part, the difference information is calculated on the basis of whether the position of the predetermined part in the input image determined by the position determining unit is an outlier using a method of robust estimation,

and wherein the detection target is a face.

11. A program for detection of a target as a face comprising computer-readable software code stored on a non-transitory recording medium that, upon execution by a processor device of a computer, causes, the computer to perform steps of:

retrieving, from a classifier storing device, information on a classifier which outputs a part likelihood of a predetermined part of a detection target calculated based on image feature values concerning the predetermined part of the detection target;

calculating, using data of an image region included in an input image, a part likelihood of the input image by using the retrieved information on the classifier;

determining a position of the predetermined part in the input image based on the calculated part likelihood;

retrieving information on a reference position of the predetermined part of the detection target stored in a reference position information storing device;

calculating difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part; and

calculating first reliability that indicates a possibility that the input image is an image showing the detection target based on a facial feature point likelihood calculated based on the image feature values,

wherein the image feature values include a difference in intensity values,

wherein, in calculating the difference information between the reference position of the predetermined part of the detection target and the determined position of the predetermined part, the difference information is calculated on the basis of whether the position of the predetermined part in the input image determined by the position determining unit is an outlier using a method of robust estimation,

and wherein the detection target is a face.