Classification of humans into multiple age categories from digital images

• US 7,319,779 B1
• Filed: 12/03/2004
• Issued: 01/15/2008
• Est. Priority Date: 12/08/2003
• Status: Active Grant

• Current Assignee: Videomining Corporation
• Original Assignee: Videomining Corporation
• Inventors: Mummareddy, Satish, Sharma, Rajeev
• Primary Examiner(s): Johns, Andrew W.
• Application Number:
  US11/004,299
• Time in Patent Office: 1,138 Days
• Field of Search: 382/100, 382/103, 382/118, 382/224, 382/226
• US Class Current: 382/118
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06V 40/172 Classification, e.g. identification
• CPC Class Codes: G06V 40/178 estimating age from face image; using age information for improving recognition

Abstract

The present invention includes a method and system for automatically extracting the multi-class age category information of a person from digital images. The system detects the face of the person(s) in an image, extracts features from the face(s), and then classifies into one of the multiple age categories. Using appearance information from the entire face gives better results as compared to currently known techniques. Moreover, the described technique can be used to extract age category information in more robust manner than currently known methods, in environments with a high degree of variability in illumination, pose and presence of occlusion. Besides use as an automated data collection system wherein given the necessary facial information as the data, the age category of the person is determined automatically, the method could also be used for targeting certain age-groups in advertisements, surveillance, human computer interaction, security enhancements and immersive computer games.

Claims

1. A method for extracting the multi-class age category information of a person from his or her facial image comprising the steps of;

a) acquiring a digital image,

b) applying a face extraction process to extract a face region of said person from said digital image,

c) processing said face region to extract face features, and

d) processing said face features obtained from said face region using a tree of binary classification systems to determine the multi-class age category,

whereby said face region is the collection of digital image pixels that encompass the face of said person in said digital image,

whereby said face features are a representation of said face region as direct pixel information or any representations obtained by transformation into other spaces by algebraic manipulation, and

whereby the multi-class age categories can be any partition of the entire age spectrum into any number of groups.

2. The method according to claim 1, wherein the step of processing said face region to extract said face features further comprises steps of applying algebraic space transformations, whereby said algebraic space transformations includes one of Principal Component Analysis, Independent Component Analysis, Non-negative Matrix Factorization or gray-scale values of the facial regions.

3. The method according to claim 1, wherein the tree structure for multi-class age classification comprises any number of binary age classifiers organized in the desired order without any restriction on the depth and fullness of the tree.

4. The method according to claim 1, wherein said binary age classification systems each comprises a single or a plurality of classifiers,

whereby the classifiers can be any pattern classification techniques, and

whereby said plurality of classifiers differ from each other by any of the processes for extracting said face features, the steps for building a classifier, the training parameters of said classifier, or a combination thereof.

5. The method according to claim 4, wherein said single or plurality of classifiers comprise a combination of two or more steps of

a. collecting data

b. training of classifier using cross-validation

c. bootstrapping to obtain the best classifier, or

d. testing the classifier.

6. The method according to claim 4, wherein said plurality of classifiers are arranged in serial,

whereby the arrangement in serial can be any sequential ordering of the plurality of classifiers, and

whereby the errors from one classifier can be detected by the next classifier in the sequence, thus improving the accuracy of the classification method.

7. The method according to claim 6, wherein the order of the classifiers could be interchanged.

8. The method according to claim 4, wherein said plurality of classifiers are arranged in parallel,

whereby the errors from one classifier can be detected by the other classifiers in parallel, thus improving the accuracy of the classification method.

9. The method according to claim 4, wherein said plurality of classifiers are arranged in a combination of serial and parallel configurations.

10. A system for extracting the multi-class age category information of a person from his or her facial image comprising;

a) means for acquiring a digital image,

b) means for applying a face extraction process to extract a face region of said person from said digital image,

c) means for processing said face region to extract face features, and

d) means for processing said face features obtained from said face region using a tree of binary classification systems to determine the multi-class age category.

11. The system according to claim 10, wherein said means for acquiring digital image comprises any digital or analog means for acquiring image,

whereby said digital or analog device can be a USB camera, a firewire camera, an infrared camera, a thermal imaging device, or any other capturing device.

12. The system according to claim 10, wherein the means for processing face features comprises one or more processing units.

13. The system according to claim 12, wherein said processing units each further comprises one or more microprocessors.