Face detection method, device and program

• US 20070172126A1
• Filed: 01/22/2007
• Published: 07/26/2007
• Est. Priority Date: 01/23/2006
• Status: Active Grant

• Current Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Original Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Inventors: Kitamura, Yoshiro
• Granted Patent: US 7,853,086 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06V 10/98 Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

To detect faces contained in an image with changing position, inclination and orientation of a face to be detected, an indicator value for each position, inclination and orientation is calculated, which indicates probability of a partial image at the position being a face image having the inclination and the orientation. Then, the partial images having the indicator values not less than a first threshold are extracted as candidates. When a predetermined evaluation value indicating reliability of a certain candidate having a predetermined inclination and a first orientation is not less than a predetermined threshold value, another candidate having the predetermined inclination and a second orientation is discriminated as being a true face. When the candidates having the indicator values not less than a second threshold are discriminated as being true faces, the second threshold is lowered for another candidate having the predetermined inclination and a second orientation for discrimination.

Claims

1. A face detection method comprising:

a face candidate extracting step for extracting face image candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating step for discriminating whether or not the face image candidates are true face images,

wherein, when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating step discriminates another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face as being a true face image.

2. The face detection method as claimed in claim 1, wherein the evaluation value is a number of the face image candidates having the first orientation of the face, and the predetermined threshold value is an integer of 2 or more.

3. The face detection method as claimed in claim 1, wherein the evaluation value is a total sum of the indicator values of the face image candidates having the first orientation of the face.

4. The face detection method as claimed in claim 1, wherein the first orientation and the second orientation are the same orientation.

5. The face detection method as claimed in claim 1, wherein the first orientation is a front orientation and the second orientation is a side or oblique orientation.

6. The face detection method as claimed in claim 4, wherein the first and second orientations comprise a plurality of different orientations.

7. A face detection method comprising:

a face candidate extracting step for extracting face candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating step for discriminating whether or not the face image candidates are true face images,

wherein the discriminating step discriminates the face image candidates having the calculated indicator values not less than a second threshold value as being true face images from all the other extracted face image candidates, and

when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating step lowers the second threshold value for discrimination of another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face.

8. The face detection method as claimed in claim 7, wherein the evaluation value is a number of the face image candidates having the first orientation of the face, and the predetermined threshold value is an integer of 2 or more.

9. The face detection method as claimed in claim 7, wherein the evaluation value is a total sum of the indicator values of the face image candidates having the first orientation of the face.

10. The face detection method as claimed in claim 7, wherein the first orientation and the second orientation are the same orientation.

11. The face detection method as claimed in claim 7, wherein the first orientation is a front orientation and the second orientation is a side or oblique orientation.

12. The face detection method as claimed in claim 10, wherein the first and second orientations comprise a plurality of different orientations.

13. A face detection device comprising:

a face candidate extracting means for extracting face image candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating means for discriminating whether or not the face image candidates are true face images,

wherein, when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating means discriminates another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face as being a true face image.

14. The face detection device as claimed in claim 13, wherein the evaluation value is a number of the face image candidates having the first orientation of the face, and the predetermined threshold value is an integer of 2 or more.

15. The face detection device as claimed in claim 13, wherein the evaluation value is a total sum of the indicator values of the face image candidates having the first orientation of the face.

16. The face detection device as claimed in claim 13, wherein the first orientation and the second orientation are the same orientation.

17. The face detection device as claimed in claim 13, wherein the first orientation is a front orientation and the second orientation is a side or oblique orientation.

18. The face detection device as claimed in claim 16, wherein the first and second orientations comprise a plurality of different orientations.

19. A face detection device comprising:

a face candidate extracting means for extracting face image candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating means for discriminating whether or not the face image candidates are true face images,

wherein the discriminating means discriminates the face image candidates having the calculated indicator values not less than a second threshold value as being true face images from all the other extracted face image candidates, and

when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating means lowers the second threshold value for discrimination of another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face.

20. The face detection device as claimed in claim 19, wherein the evaluation value is a number of the face image candidates having the first orientation of the face, and the predetermined threshold value is an integer of 2 or more.

21. The face detection device as claimed in claim 19, wherein the evaluation value is a total sum of the indicator values of the face image candidates having the first orientation of the face.

22. The face detection device as claimed in claim 19, wherein the first orientation and the second orientation are the same orientation.

23. The face detection device as claimed in claim 19, wherein the first orientation is a front orientation and the second orientation is a side or oblique orientation.

24. The face detection device as claimed in claim 22, wherein the first and second orientations comprise a plurality of different orientations.

25. A program for causing a computer to operate as a face detection device by causing the computer to operate as:

a face candidate extracting means for extracting face image candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating means for discriminating whether or not the face image candidates are true face images,

wherein, when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating means discriminates another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face as being a true face image.

26. A program for causing a computer to operate as a face detection device by causing the computer to operate as:

a face candidate extracting means for extracting face image candidates by changing position, inclination and orientation of a face to be detected in an inputted image, calculating, for each of the position, inclination and orientation of the face to be detected, an indicator value indicating a probability of a partial image at the position being a face image containing a face having the inclination and the orientation, and extracting all the partial images having the indicator values not less than a first threshold value as face image candidates; and

a discriminating means for discriminating whether or not the face image candidates are true face images,

wherein the discriminating means discriminates the face image candidates having the calculated indicator values not less than a second threshold value as being true face images from all the other extracted face image candidates, and

when a predetermined evaluation value indicating reliability of a certain face image candidate, among the face image candidates, having a predetermined inclination and a first orientation of the face is not less than a predetermined threshold value, the discriminating means lowers the second threshold value for discrimination of another face image candidate, among the face image candidates, having substantially the same inclination as the predetermined inclination and a second orientation of the face.