Face or Other Object Detection Including Template Matching

• US 20120106790A1
• Filed: 10/26/2011
• Published: 05/03/2012
• Est. Priority Date: 10/26/2010
• Status: Active Grant

• Current Assignee: FotoNation Limited (Adeia Inc.)
• Original Assignee: DigitalOptics Corporation Europe Limited (Adeia Inc.)
• Inventors: Petrescu, Stefan, Bigioi, Petronel, Corcoran, Peter, Gangea, Mihnea, Sultana, Bogdan, Nicolau, Radu, Ursachi, Vlad Ionut, Zaharia, Corneliu, Fulop, Szabolcs
• Granted Patent: US 8,995,715 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06V 10/60 relating to illumination properties, e.g. using a reflectance or lighting model
• CPC Class Codes: G06V 10/75 Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
• CPC Class Codes: G06V 10/94 Hardware or software architectures specially adapted for image or video understanding
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation
• CPC Class Codes: G06V 40/164 using holistic features

Abstract

A template matching module is configured to program a processor to apply multiple differently-tuned object detection classifier sets in parallel to a digital image to determine one or more of an object type, configuration, orientation, pose or illumination condition, and to dynamically switch between object detection templates to match a determined object type, configuration, orientation, pose, blur, exposure and/or directional illumination condition.

Claims

1. A digital image acquisition device, comprising:

an optoelectonic system for acquiring a digital image;

a processor; and

a template matching module configured to program the processor to apply multiple differently-tuned object detection classifier sets in parallel to the digital image to determine one or more of an object type, configuration, orientation, pose or illumination condition, and to dynamically switch between object detection templates to match a determined object type, configuration, orientation, pose, blur, exposure or directional illumination condition, or combinations thereof.

2. The device of claim 1, wherein one or more of the multiple differently-tuned object detection classifier sets comprise a subset of a full classifier set including an initial 3-8 classifiers of said full set.

3. The device of claim 1, wherein the multiple differently-tuned classifier sets comprise classifier sets tuned to detect multiple different selected objects or to recognize faces of multiple different specific persons, or both.

4. The device of claim 1, wherein the template matching module comprises a dedicated hardware component.

5. The device of claim 1, wherein the template matching module is configured to program the processor to dynamically switch between said templates based on a frame to frame analysis.

6. The device of claim 1, wherein the template matching module is configured to program the processor to dynamically switch between said templates selectively on a local scale.

7. The device of claim 1, further comprising a hardware acceleration block configured to receive data acquired from the optoelectronic system and to provide one or more hardware acceleration maps to a memory, and the template matching module is further configured to retrieve luminance, grayscale or Y data, or combinations thereof, from the memory.

8. The device of claim 1, wherein at least one of the multiple differently-tuned object detection classifier sets comprise a face detection classifier set.

9. A template matching module configured to program a processor to apply multiple differently-tuned object detection classifier sets in parallel to a digital image to determine one or more of an object type, configuration, orientation, pose or illumination condition, and to dynamically switch between object detection templates to match a determined object type, configuration, orientation, pose, blur, exposure or directional illumination condition, or combinations thereof.

10. The module of claim 9, wherein one or more of the multiple differently-tuned object detection classifier sets comprise a subset of a full classifier set including an initial 3-8 classifiers of said full set.

11. The module of claim 9, wherein the multiple differently-tuned classifier sets comprise classifier sets tuned to detect multiple different selected objects or to recognize faces of multiple different specific persons, or both.

12. The module of claim 9, wherein the template matching module comprises a dedicated hardware component.

13. The module of claim 9, wherein the template matching module is configured to program the processor to dynamically switch between said templates based on a frame to frame analysis.

14. The module of claim 9, wherein the template matching module is configured to program the processor to dynamically switch between said templates selectively on a local scale.

15. The module of claim 9, wherein at least one of the multiple differently-tuned object detection classifier sets comprise a face detection classifier set.

16. An object detection system, comprising:

a template matching module configured to program a processor to apply multiple differently-tuned object detection classifier sets in parallel to a digital image to determine one or more of an object type, configuration, orientation, pose or illumination condition, and to dynamically switch between object detection templates to match a determined object type, configuration, orientation, pose, blur, exposure or directional illumination condition, or combinations thereof; and

a hardware acceleration block configured to receive acquired image data and to provide one or more hardware acceleration maps to a memory, from which the template matching module is further configured to retrieve luminance, grayscale or Y data, or combinations thereof.

17. The system of claim 116, wherein at least one of the multiple differently-tuned object detection classifier sets comprise a face detection classifier set.

18. A method of object detection in an acquired digital image, comprising:

acquiring a digital image;

applying multiple differently-tuned object detection classifier sets in parallel to the digital image to determine one or more of an object type, configuration, orientation, pose or illumination condition; and

dynamically switching between object detection templates to match a determined object type, configuration, orientation, pose, blur, exposure or directional illumination condition, or combinations thereof.

19. The method of claim 18, wherein one or more of the multiple differently-tuned object detection classifier sets comprise a subset of a full classifier set including an initial 3-8 classifiers of said full set.

20. The method of claim 18, wherein the multiple differently-tuned classifier sets comprise classifier sets tuned to detect multiple different selected objects or to recognize faces of multiple different specific persons, or both.

21. The method of claim 18, wherein at least one of the multiple differently-tuned object detection classifier sets comprise a face detection classifier set.

22. The method of claim 18, wherein the template matching module comprises a dedicated hardware component.

23. The method of claim 18, further comprising dynamically switching between said templates based on a frame to frame analysis.

24. The method of claim 18, further comprising dynamically switching between said templates selectively on a local scale.

25. The method of claim 18, further comprising providing acquired image data and one or more hardware acceleration maps to a memory, and retrieving from the memory luminance, grayscale or Y data, or combinations thereof.