Method and apparatus for estimating position of head, computer readable storage medium thereof

• US 9,349,038 B2
• Filed: 05/07/2013
• Issued: 05/24/2016
• Est. Priority Date: 09/11/2012
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Puri, Anant Vidur, Parihar, Abhinav
• Primary Examiner(s): Sherali, Ishrat I
• Application Number:
  US13/888,612
• Publication Number:
  US 20140072234A1
• Time in Patent Office: 1,113 Days
• Field of Search: 382/201, 382/103
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/77 using statistical methods
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

A computer implemented method and an apparatus for estimating head position of an individual are described. The method includes cropping an image object in an image. The image object may include a visual representation of a head of at least one individual. The method further includes abstracting one or more features from the image object. Subsequently,

• • position of the head of the at least one individual in the image object is determined using the one or more features and a predefined statistical model. The predefined statistical model may include a value of each of the pixels of one or more stored images representing variety of head poses and ranges of head poses of the at least one individual.

Claims

1. A computer-implemented head position estimating method comprising:

cropping an image object in an image, wherein the image object comprises a visual representation of a head of at least one individual;

abstracting one or more features from the image object; and

determining a position of the head of the at least one individual in the image object, by way of a processor, using the one or more abstracted features and a predefined statistical model, wherein the determining the position of the head comprises calculating a probability that each of the pixels in the image object belongs to a set of poses of a plurality of sets of poses in the predefined statistical model, and determining the position of the head based on the calculated probability.

2. The computer-implemented method as recited in claim 1, wherein the predefined statistical model comprises a value of each of the pixels of one or more stored images that represent a variety of head poses and ranges of head poses of the at least one individual.

3. The computer-implemented method as recited in claim 1, wherein the determining the position of the head further comprises:

computing a log likelihood sum for each of the plurality of sets of poses based on the probability values of the pixels in the image object; and

determining one set of the plurality of sets of poses having a maximum log likelihood sum as the position of the head in the image object.

4. The computer-implemented method as recited in claim 1, wherein the one or more features comprises at least one of eyes, nose, hair, ears, cheek, chin and lips.

5. The computer-implemented method as recited in claim 1, further comprising capturing an image which visually represents the head of the at least one individual.

6. The computer-implemented method as recited in claim 1, further comprising resizing the image object to a pre-defined size upon the abstracting of the one or more features from the image object.

7. The computer-implemented method as recited in claim 1, wherein the abstracting the one or more features from the image object is performed using a non-photorealistic rendering technique.

8. The computer-implemented method as recited in claim 1, wherein the abstracting the one or more features from the image object comprises:

identifying a color of each of the pixels in the image object; and

segmenting pixels of the image object into groups of similar colors for identifying the one or more features.

9. An apparatus comprising:

a processor configured to control one or more processor-executable modules;

a cropping module among the one or more processor-executable modules, the cropping module being configured for cropping an image object in an image, wherein the image object comprises a visual representation of a head of at least one individual;

an abstraction module among the one or more processor-executable modules, the abstraction module being configured for abstracting one or more features from the image object; and

a head position determining module among the one or more processor-executable modules, the head positioning determining module being configured for determining a position of the head of the at least one individual in the image object using the one or more features abstracted by the abstraction module and a predefined statistical model, wherein the determining the position of the head comprises calculating a probability that each of the pixels in the image object belongs to a set of poses of a plurality of sets of poses in the predetermined statistical model, and determining the position of the head based on the calculated probability.

10. The apparatus of claim 9, wherein the predefined statistical model comprises a value of each of the pixels of one or more stored images of a variety of positions of the at least one individual.

11. The apparatus as recited in claim 9, wherein the determining the position of the head further comprises:

computing a log likelihood sum for each of the plurality of sets of poses based on the probability values of the pixels in the image object; and

determining one set of the plurality of sets of poses having a maximum log likelihood sum as position of the head in the image object.

12. The apparatus as recited in claim 11, wherein in the abstracting the one or more features from the image object, the abstraction module is configured for:

identifying a color of each of the pixels in the image object; and

segmenting pixels in the image object into groups, wherein each of the groups of pixels belong to a specific color.

13. A non-transitory computer readable storage medium having instructions stored therein, that when executed by a computing device, cause the computing to perform a method, the method comprising:

cropping an image object in an image, wherein the image object comprises a visual representation of a head of at least one individual;

abstracting one or more features from the image object; and

determining a position of the head of the at least one individual in the image object using the one or more abstracted features and a predefined statistical model, wherein the determining the position of the head comprises calculating a probability that each of the pixels in the image object belongs to a set of poses of a plurality of sets of poses in the predetermined statistical model, and determining the position of the head based on the calculated probability.

14. The storage medium according to claim 13, wherein the predefined statistical model comprises a value of each of the pixels of one or more stored images that represent a variety of head poses and ranges of head poses of the at least one individual.

15. The storage medium according to claim 13, wherein the determining the position of the head further comprises:

computing a log likelihood sum for each of the plurality of sets of poses based on the probability values of the pixels in the image object; and

determining one set of the plurality of sets of poses having a maximum log likelihood sum as position of the head in the image object.

16. The storage medium according to claim 13, wherein the one or more features comprises at least one of eyes, nose, hair, ears, cheek, chin and lips.

17. The storage medium according to claim 13, the method further comprising capturing an image which visually represents the head of the at least one individual.

18. The storage medium according to claim 13, the method further comprising resizing the image object to a pre-defined size upon the abstracting of the one or more features from the image object.

19. The storage medium according to claim 13, wherein the abstracting the one or more features from the image object is performed using a non-photorealistic rendering technique.

20. The storage medium according to claim 13, wherein abstracting the one or more features from the image object comprises:

identifying color of each of the pixels in the image object; and

segmenting pixels of the image object into groups of similar colors for identifying the one or more features.

21. A head position estimating apparatus comprising:

a processor configured to control one or more processor-executable modules;

an abstraction module among the one or more processor-executable modules, the abstraction module being configured to abstract one or more features of an image object comprising a visual representation of a head of at least one person; and

a head position determining module among the one or more processor-executable modules, the head position determining module being configured to determine a position of the head of the at least one person in the image object using the one or more features abstracted by the abstraction module and a predefined statistical model, wherein the determining the position of the head comprises calculating a probability that each of the pixels in the image object belongs to a set of poses of a plurality of sets of poses in the predefined statistical model, and determining the position of the head based on the calculated probability.

22. The head position estimating apparatus as recited in claim 21, further comprising a cropping module configured to crop the image object in the image, and wherein the abstraction module is further configured to abstract the one or more features from the cropped image object.