Method for detecting a human face and an apparatus of the same

• US 6,885,760 B2
• Filed: 01/29/2001
• Issued: 04/26/2005
• Est. Priority Date: 02/01/2000
• Status: Expired due to Term

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Souma, Masaki, Itoh, Ryuta, Nagao, Kenji, Yamada, Shin
• Primary Examiner(s): Bali, Vikkram
• Application Number:
  US09/770,639
• Publication Number:
  US 20010036298A1
• Time in Patent Office: 1,548 Days
• Field of Search: 382/103, 382/115, 382/117, 382/118, 382/170, 382/171, 382/224, 235/380, 235/382, 235/382.5, 340/5.53, 340/5.83, 902 2- 6, 351/204, 514/210.1
• US Class Current: 382/118
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation
• CPC Class Codes: G06V 40/178 estimating age from face image; using age information for improving recognition

Abstract

In an apparatus for detecting a human face from an input image, candidate areas for an eye and candidate areas for an interval between eyes are detected from the image, an area of the interval between eyes is determined from the positional relationship between the candidate areas for an eye and candidate area for an interval between eyes, and a face area is determined so that the face area includes the area of the interval between eyes.

Claims

1. A detecting method for detecting a human face from an input image, comprising:

detecting at least first and second candidate eye areas for first and second eyes and a plurality of candidate interval areas for an interval between the first and second eyes based on luminance characteristics of the input image;

determining whether a candidate interval area among the plurality of candidate interval areas is between the first and second candidate eye areas;

designating the candidate interval area as the interval between the first and second eyes when the candidate interval area is determined to be between the first and second candidate eye areas; and

determining a face area based on the designated interval between the first and second eyes.

2. The method according to claim 1, wherein the first and second candidate eye areas are detected from the luminance characteristics corresponding to a first plurality of portions of the input image having a substantially horizontal alignment and a second plurality of portions of the input image having a substantially vertical alignment, and

wherein there are fewer of the second plurality of portions than of the first plurality of portions.

3. The method according to claim 1, wherein the candidate interval areas are detected based on the luminance characteristics corresponding to a plurality of areas of the input image having fewer substantially horizontal portions and substantially vertical portions than a threshold amount.

4. The method according to claim 1, further comprising:

dividing at least a part of the input image into a plurality of sub areas, including the first and second candidate eye areas and the candidate interval areas.

5. The method according to claim 4, wherein the candidate eye areas are detected using edge information of a plurality of pixels in the sub areas.

6. The method according to claim 4, wherein the candidate eve areas are detected by obtaining a luminance histogram of a plurality of pixels in each of the sub areas, and comparing the luminance histogram between adjacent sub areas.

7. The method according to claim 4, wherein the candidate interval areas are detected using edge information of a plurality of pixels in the sub areas.

8. The method according to claim 7, wherein the candidate interval areas are detected using the edge information of the pixels in the sub areas adjacent each other in a vertical direction.

9. The method according claim 4, wherein the candidate interval areas are detected by obtaining a luminance histogram of a plurality of pixels in each of the sub areas, and comparing the luminance histogram between adjacent small areas.

10. The method according to claim 1, wherein information of a face in the face area is detected.

11. The method according to claim 10, wherein the information of the face includes at least one of a sex and an age.

12. A moving picture retrieval method, comprising:

detecting the face area using the method according to claim 1 from a frame picture of a moving picture;

storing information of a face in each area; and

retrieving a moving picture using the information on the face.

13. A face identifying method for identifying a face, comprising:

detecting the face area using the method according to claim 1; and

identifying a face using a feature vector of a face in each area and feature vectors of faces stored in a database.

14. A detecting method for detecting a human face from an input image, comprising:

detecting a plurality of candidate eye areas for first and second eyes based on luminance characteristics of the input image;

detecting an interval between the first and second eyes;

examining one of luminance characteristics and shape characteristics of the candidate eye areas for the first and second eyes to determine a position of a right eye and a position of a left eye; and

determining a face area including the position of the right eye and the position of the left eye.

15. A detecting apparatus for detecting a human face from an input image, comprising:

an eye candidate detecting section configured to detect an interval between a plurality of candidate eye areas for first and second eyes based on a plurality of luminance characteristics of the input image;

an interval detecting section configured to detect an interval between the first and second eyes;

an eye determining section configured to examine one of luminance characteristics and shape characteristics of the candidate eye areas for the first and second eyes detected by said eye candidate detecting section to determine a position of a right eye and a position of a left eye; and

a face determining section configured to determine a face area including the position of the right eye and the position of the left eye determined by said eye determining section.

16. A computer readable storage medium storing a detecting program for detecting a human face from an image, said detecting program configured to cause a computer to perform at least:

detecting at least first and second candidate eye areas for first and second eyes and a plurality of candidate interval areas for an interval between the first and second eyes based on luminance characteristics of the input image;

determining whether a candidate interval area among the plurality of candidate interval areas is between the first and second candidate eye areas;

designating the candidate interval area as the interval between the first and second eyes when the candidate interval area is determined to be between the first and second candidate eye areas; and

determining a face area based on the designated interval between the first and second eyes.

17. A computer readable storage medium storing a detecting program for detecting a human face from an image, said program configured to cause a computer to perform at least:

detecting a plurality of candidate eye areas for first and second eyes based on a plurality of luminance characteristics of the image;

detecting an interval between the first and second eyes;

examining one of luminance characteristics and shape characteristics of the candidate areas for both eyes to determine a position of a right eye and a position of a left eye; and

determining a face area including the position of the right eye and the position of the left eye.

18. A computer readable storage medium storing a motion picture retrieval program, said motion picture retrieval program configured to cause a computer to perform at least:

detecting at least first and second candidate eye areas for first and second eyes and a plurality of candidate interval areas for an interval between the first and second eyes based on a plurality of luminance characteristics of a frame picture of a motion picture;

determining whether a candidate interval area among the plurality of candidate interval areas is between the first and second candidate eye areas;

designating the candidate interval area as the interval between the first and second eyes when the candidate interval area is determined to be between the first and second candidate eye areas;

determining a face area based on the designated interval between the first and second eyes;

storing information of a face in the face area; and

retrieving a motion picture using the information of the face.

19. A computer readable storage medium storing a face identifying program, said program configured to cause a computer to perform at least:

detecting at least first and second candidate eye areas for first and second eyes and a plurality of candidate interval areas for an interval between the first and second eyes based on a plurality of luminance characteristics of an image;

determining whether a candidate interval area among the plurality of candidate interval areas is between the first and second candidate eye areas;

designating the candidate interval area as the interval between the first and second eyes when the candidate interval area is determined to be between the first and second candidate eye areas;

determining a face area based on the designated interval between the first and second eyes; and

identifying a face using a feature vector of the face extracted from the face area compared with a plurality of feature vectors of faces stored in a database.