FACE DETECTION METHOD AND DIGITAL CAMERA

• US 20080310726A1
• Filed: 06/13/2008
• Published: 12/18/2008
• Est. Priority Date: 06/18/2007
• Status: Active Grant

• Current Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Original Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Inventors: Kawada, Yukihiro
• Granted Patent: US 8,005,271 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches
• CPC Class Codes: G06V 40/164 using holistic features

Abstract

A face detection method is provided, including: classifying into levels time-wise continuously captured images by increasing/reducing the total number of pixels; selecting sequentially and reading out image data for all of the levels using read-out units of the same size of pixels, or a smaller size of pixels, as those of the image with the smallest size of pixels; carrying out face detection processing by extracting candidate levels in which face image data is present, based on the read-out image data for each of the levels; and, when repeating the face detection processing, setting the number of candidate levels for face detection processing from the second time onward as less than the total number of levels. A digital camera incorporating the face detection method is also disclosed.

Claims

1. A face detection method comprising:

classifying a plurality of images, generated by enlargement or reduction of time-wise successively captured images according to a plurality of specific multipliers, into levels by image size of the images;

sequentially reading out at one read-out unit at a time the images classified into each of the levels, using as the read-out unit a size that is smaller than the smallest image size from the image sizes of the classified images;

carrying out face detection processing to detect images in which it is determined that a face is present in portions of the image divided up into the read-out units, wherein when the face detection processing is repeatedly carried out to the time-wise successively captured images the number of candidate levels of the images used for the face detection processing for the next image is set based on the level(s) of image size of images in which it has been determined there is a face present so that the number of candidate levels is less than the total number of levels.

2. The face detection method according to claim 1, wherein the settable number of candidate levels is set based on the image size(s) in which it has been determined that a face is present the previous time face detection processing was carried out.

3. The face detection method according to claim 1, further comprising:

additionally setting a function for discriminating using an out-plane angle, representing the left-right facing direction of a face, as a parameter for subjecting to face detection processing, wherein the number of types of the out-plane angle to be subjected to face detection processing is set to be fewer as the image size of the classified images increases.

4. The face detection method according to claim 2, further comprising:

additionally setting a function for discriminating using an out-plane angle, representing the left-right facing direction of a face, as a parameter for subjecting to face detection processing, wherein the number of types of the out-plane angle to be subjected to face detection processing is set to be fewer as the image size of the classified images increases.

5. The face detection method according to claim 3, wherein the face detection processing is carried out within a region set in advance for each of the levels of image, the region being set according to the out-plane angle(s) in which a face has been determined to be present by the face detection processing the previous time.

6. The face detection method according to claim 4, wherein the face detection processing is carried out within a region set in advance for each of the levels of image, the region being set according to the out-plane angle(s) in which a face has been determined to be present by the face detection processing the previous time.

7. The face detection method according to claim 1, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

8. The face detection method according to claim 2, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

9. The face detection method according to claim 3, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

10. The face detection method according to claim 4, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

11. The face detection method according to claim 5, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

12. The face detection method according to claim 6, further comprising:

additionally setting a function for discriminating using an in-plane angle, representing the degree of angular rotation of a face to the upright direction, as a parameter for subjecting to face detection processing, wherein when the face detection processing is carried out repeatedly the number of patterns of the in-plane angle to be used in the face detection processing for the next image is set to be fewer than the total number of the patterns, and the number of patterns are set so as to belong to a specific angular range around the in-plane angle in which a face was detected to be present the previous time.

13. The face detection method according to claim 1, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

14. The face detection method according to claim 2, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

15. The face detection method according to claim 3, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

16. The face detection method according to claim 4, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

17. The face detection method according to claim 5, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

18. The face detection method according to claim 6, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

19. The face detection method according to claim 7, wherein, the face detection processing is carried out within a region set in advance for each of the levels of image according to the image size of the images for each of the levels.

20. A digital camera incorporating the face detection method of claim 1.