Method and apparatus for red-eye detection in an acquired digital image

• US 7,587,085 B2
• Filed: 09/21/2005
• Issued: 09/08/2009
• Est. Priority Date: 10/28/2004
• Status: Active Grant

• Current Assignee: FotoNation Limited (Adeia Inc.)
• Original Assignee: FotoNation Vision Limited (Adeia Inc.)
• Inventors: Nanu, Florin, Steinberg, Eran, Prilutsky, Yury, Bigioi, Petronel
• Primary Examiner(s): Do; Anh Hong
• Application Number:
  US11/233,513
• Publication Number:
  US 20060120599A1
• Time in Patent Office: 1,448 Days
• Field of Search: 382/167, 382/117, 382/263, 382/275, 382/165, 382/118, 382/164, 382/103, 348/223.1, 348/252, 348/341, 348/E5.042, 358/518
• US Class Current: 382/167
• CPC Class Codes: G06T 2207/20008 Globally adaptive
• CPC Class Codes: G06T 2207/30216 Redeye defect
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement
• CPC Class Codes: G06V 40/193 Preprocessing; Feature extraction

Abstract

A method for red-eye detection in an acquired digital image comprises acquiring a first image and analyzing the first acquired image to provide a plurality of characteristics indicative of image quality. The process then determines if one or more corrective processes can be beneficially applied to the first acquired image according to the characteristics. Any such corrective processes are then applied to the first acquired image. Red-eye defects are then detected in a second acquired image using the corrected first acquired image. Defect detection can comprise applying a chain of one or more red-eye filters to the first acquired image. In this case, prior to the detecting step, it is determined if the red-eye filter chain can be adapted in accordance with the plurality of characteristics; and the red-eye filter is adapted accordingly.

Claims

1. A method for red-eye detection in an acquired digital image comprising, using a processor and a digital storage medium having code embedded therein for programming the processor to perform the method, which includes:

a) acquiring a first image;

b) analyzing the first acquired image to detect one or more eye color defect characteristics;

c) determining if one or more corrective processes can be beneficially applied to said first acquired image according to said eye color defect characteristics;

d) applying at least one of such corrective processes to said first acquired image; and

e) detecting red-eye defects in a second acquired image using said corrected first acquired image.

2. A method according to claim 1 wherein said detecting comprises applying a chain of one or more red-eye filters to said first acquired image;

and further comprising, prior to said detecting;

f) determining if said red-eye filter chain can be adapted in accordance with said plurality of characteristics; and

g) adapting said red-eye filter chain accordingly.

3. A method according to claim 2 in which said adapting comprises providing an altered set of parameters for one or more filters of said filter chain.

4. A method according to claim 2 in which said adapting comprises re-ordering a sequence in which said filters are applied to said first acquired image.

5. A method according to claim 2 in which more than one of said filters are applied simultaneously.

6. A method according to claim 2 in which said adapting comprises determining which filters are to be applied in said chain.

7. A method according to claim 1 further comprising correcting a third acquired image based on any defects detected in said detecting.

8. A method according to claim 7 wherein said first second and third images are the same image.

9. A method according to claim 7 wherein at least one of said first and second images is a sub-sampled copy of said third image.

10. A method according to claim 1 wherein said first and second acquired images are the same image.

11. A method according to claim 1 wherein said first and second acquired images are different images.

12. A method according to claim 11 wherein said first acquired image is a sub-sampled copy of an acquired image.

13. A method according to claim 11 wherein said first acquired image is a sub-sampled copy of the second image.

14. A method according to claim 1 further comprising:

prior to analyzing the first acquired image, processing said second image based on the settings of a device used to acquire the second image.

15. A method according to claim 1 wherein said analyzing comprises determining one or more of:

a degree of blur;

a degree of dust contamination;

color balance;

white balance;

a gamma correction which might be applied;

texture characteristics;

noise characteristics; and

characteristics of regions;

in the first acquired image.

16. A method according to claim 1 wherein the analyzing comprises recognizing one or more faces in the first acquired image.

17. A method according to claim 16, further comprising recognizing a same face in a second acquired image as in said first acquired image, and applying one or more same corrective processes to said second image as said first image.

18. A method according to claim 1 wherein said corrective processes comprise one or more of:

(i) contrast normalization and image sharpening;

(ii) image color adjustment and tone scaling;

(iii) exposure adjustment and digital fill flash;

(iv) brightness adjustment with color space matching;

(v) image auto-gamma determination with image enhancement;

(v) image enhancement; and

(vi) face based image enhancement.

19. A method according to claim 2 wherein said red-eye filter chain comprises:

(i) a pixel locator and segmentor;

(ii) a shape analyser;

(iii) a falsing analyser; and

(iv) a pixel modifier.

20. A method according to claim 19 wherein said pixel locator and segmentor includes a pixel transformer.

21. A method as claimed in claim 1 comprising:

responsive to determining that corrective processing is necessary but cannot be beneficially applied to said first acquired image according to said characteristics, disabling said detecting red-eye defects and providing an indication of such to a user.

22. A method as claimed in claim 7 wherein said applying a corrective process includes interacting with a user to determine the corrections to be made to said third image.

23. A method according to claim 1 wherein the applying a corrective process includes performing a color space transformation.

24. A method as claimed in claim 1 in which one or more of (a) to (e) are performed in a digital camera.

25. A method as claimed in claim 1 comprising determining a sequence in which more than one corrective process can be beneficially applied to said first acquired image according to said characteristics.

26. A method according to claim 1 wherein the analyzing comprises recognizing one or more types of faces in the first acquired image.

27. A method according to claim 16, further comprising recognizing a same type of face in a second acquired image as in said first acquired image, and applying one or more same corrective processes to said second image as said first image.

28. A computer-readable storage medium containing a set of instructions which when executed on a digital image processing device program a processor to perform a method for red-eye detection in an acquired digital image, wherein the method comprises:

a) acquiring a first image;

b) analyzing the first acquired image to detect one or more eye color defect characteristics;

c) determining if one or more corrective processes can be beneficially applied to said first acquired image according to said eye color detect characteristics; and

d) applying at least one of such corrective processes to said first acquired image; and

e) detecting red-eye defects in a second acquired image using said corrected first acquired image.

29. A digital image processing device comprising a processor and a processor-readable storage medium having embedded therein processor-readable code for programming the processor to perform a method for red-eye detection in an acquired digital image, wherein the method comprises:

a) acquiring a first image;

b) analyzing the first acquired image to detect one or more eye color defect characteristics;

c) determining if one or more corrective processes can be beneficially applied to said first acquired image according to said eye color defect characteristics; and

d) applying at least one of such corrective processes to said first acquired image; and

e) detecting red-eye defects in a second acquired image using said corrected first acquired image.

30. A method for red-eye detection in an acquired digital image comprising, using a processor and a digital storage medium having code embedded therein for programming the processor to perform the method, which includes:

a) acquiring a first image;

b) analyzing the first acquired image to detect a suboptimal exposure characteristic of the image;

c) determining if a chain of one or more red-eye filters can be adapted in accordance with said suboptimal exposure characteristic;

d) adapting said red-eye filter chain; and

e) applying said filter chain to detect red-eye defects in a second acquired image.

31. A method according to claim 30 further comprising:

f) determining if one or more corrective processes can be beneficially applied to said first acquired image according to said characteristics; and

g) applying any such corrective processes to said first acquired image.

32. A method according to claim 30 in which said adapting comprises providing an altered set of parameters for one or more filters of said filter chain.

33. A method according to claim 30 in which said adapting comprises re-ordering a sequence in which said filters are applied to said first acquired image.

34. A method according to claim 30 in which more than one of said filters are applied simultaneously.

35. A method according to claim 30 in which said adapting comprises determining which filters are to be applied in said chain.

36. A method according to claim 30 further comprising correcting a third acquired image based on one or more defects detected in said detecting.

37. A method according to claim 36 wherein said first second and third images comprise a same image.

38. A method according to claim 36 wherein at least one of said first and second images is a sub-sampled copy of said third image.

39. A method according to claim 30 wherein said first and second acquired images comprise a same image.

40. A method according to claim 30 wherein said first and second acquired images comprise different images.

41. A method according to claim 40 wherein said first acquired image comprises a sub-sampled copy of an acquired image.

42. A method according to claim 40 wherein said first acquired image comprises a sub-sampled copy of the second image.

43. A method according to claim 30 further comprising:

prior to analyzing the first acquired image, processing said second image based on the settings of a device used to acquire the second image.

44. A method according to claim 30 wherein said analysing step comprises determining one or more of:

a degree of blur a degree of dust contamination;

color balance;

white balance;

a gamma correction which might be applied;

texture characteristics;

noise characteristics; and

characteristics of regions;

in the first acquired image.

45. A method according to claim 30 wherein said analyzing comprises recognizing one or more faces in the first acquired image.

46. A method according to claim 45, further comprising recognizing a same face in said second acquired image as in said first acquired image, and wherein said applying said filter chain comprises applying a same filter chain to said same face of said second acquired image as in said first acquired image.

47. A method according to claim 31 wherein said corrective processes comprise one or more of:

(i) contrast normalization and image sharpening;

(ii) image color adjustment and tone scaling;

(iii) exposure adjustment and digital fill flash;

(iv) brightness adjustment with color space matching;

(v) image auto-gamma determination with image enhancement;

(v) image enhancement; and

(vi) face based image enhancement.

48. A method according to claim 30 wherein said red-eye filter chain comprises:

(i) a pixel locator and segmentor;

(ii) a shape analyser;

(iii) a falsing analyser; and

(iv) a pixel modifier.

49. A method according to claim 48 wherein said pixel locator and segmentor includes a pixel transformer.

50. A method as claimed in claim 31 comprising responsive to determining that corrective processing is necessary but cannot be beneficially applied to said first acquired image according to said characteristics, disabling said detecting red-eye defects and providing an indication of such to a user.

51. A method as claimed in claim 36 wherein said applying a corrective process includes interacting with a user to determine the corrections to be made to said third image.

52. A method according to claim 31 wherein the applying a corrective process includes performing a color space transformation.

53. A method as claimed in claim 30 in which one or more of (a) to (e) are performed in a digital camera.

54. A method as claimed in claim 31 comprising determining a sequence in which more than one corrective process can be beneficially applied to said first acquired image according to said characteristics.

55. A method according to claim 30 wherein said analyzing comprises recognizing one or more types of faces in the first acquired image.

56. A method according to claim 45, further comprising recognizing a same type of face in said second acquired image as in said first acquired image, and wherein said applying said filter chain comprises applying a same filter chain to said same type of face of said second acquired image as in said first acquired image.

57. A computer-readable storage medium containing a set of instructions embedded therein which, when executed on a digital image processing device, that comprises a processor, programs the processor to perform a method for red-eye detection in an acquired digital image, wherein the method comprises:

a) acquiring a first image;

b) analyzing the first acquired image to detect a suboptimal exposure characteristic of the image;

c) determining if a chain of one or more red-eye filters can be adapted in accordance with said suboptimal exposure characteristic;

d) adapting said red-eye filter chain; and

e) applying said filter chain to detect red-eye defects in a second acquired image.

58. A digital image acquisition and processing device comprising a lens, an image sensor, a processor and a processor-readable storage medium having embedded therein processor-readable code for programming the processor to perform a method for red-eye detection in an acquired digital image, wherein the method comprises:

a) acquiring a first image;

b) analyzing the first acquired image to detect a suboptimal exposure characteristic of the image;

c) determining if a chain of one or more red-eye filters can be adapted in accordance with said suboptimal exposure characteristic;

d) adapting said red-eye filter chain; and

e) applying said filter chain to detect red-eye defects in a second acquired image.