Tilt correction of electronic images

• US 7,305,146 B2
• Filed: 05/20/2002
• Issued: 12/04/2007
• Est. Priority Date: 06/30/2001
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Cheatle, Stephen Philip
• Primary Examiner(s): Alavi; Amir
• Application Number:
  US10/343,156
• Publication Number:
  US 20030152291A1
• Time in Patent Office: 2,024 Days
• Field of Search: 382/118, 382/162, 382/254, 382/276, 382/278, 382/282, 382/289, 382/291, 382/293, 382/190, 382/195, 382/275, 382/296, 358/429, 358/474, 358486-488, 358/490, 358/496, 358/497
• US Class Current: 382/296
• CPC Class Codes: G06T 2207/20132 Image cropping
• CPC Class Codes: G06T 3/608 by skew deformation, e.g. two-pass or three-pass rotation
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: H04N 1/3873 defined only by a limited number of coordinate points or parameters, e.g. corners, centre; for trimming
• CPC Class Codes: H04N 1/3878 Skew detection or correction

Abstract

A method of correcting the tilt, or rotation, of a casually captured image is described. Having corrected the rotation of the original image, the image is cropped by determining a crop boundary by applying one or more rules of composition to the image. The resulting image is more satisfactorily composed compared with prior art methods.

Claims

1. A method of correcting tilt of an electronic image, the method comprising:

rotating an original, tilted image by a correcting angle of rotation;

analyzing different image features within the original image to determine a subject matter of the original image; and

automatically cropping the resulting rotated image in accordance with at least one rule selected from a predefined set of image composition rules, wherein the rule is based on the subject matter of the original image, such that the crop boundary is contained within the original image.

2. A method of correcting tilt of an electronic image according to claim 1, wherein said image composition rules comprise:

identifying one or more elements of interest within said image;

selecting one or more of said identified elements to be included within the cropped image; and

generating at least one crop boundary so as to include said selected elements of interest.

3. A method of correcting tilt of an electronic image as claimed in claim 2, wherein the crop boundaries are calculated such that the area defined by the boundaries is determined on the basis of the image features in the original image.

4. A method as claimed in claim 1, wherein the crop boundary is rectangular and has one axis in a notionally vertical direction.

5. A method of correcting tilt of an electronic image according to claim 1, wherein said electronic image has a predefined aspect ratio and said rotated image is cropped such that the cropped image has said predefined aspect ratio.

6. A method of correcting tilt of an electronic image according to claim 1, wherein said crop boundary is smaller than the largest crop boundary that can be fitted within said original tilted image.

7. A method of correcting tilt of an electronic image according to claim 1, wherein said correcting angle of rotation is determined at the time of capturing said original image.

8. A method of correcting tilt of an electronic image according to claim 1, wherein said correcting angle of rotation is defined by a manual input.

9. A method of correcting tilt of an electronic image according to claim 1, wherein said correcting angle of rotation is determined by analysis of said image.

10. A method of correcting tilt of an electronic image according to claim 9, wherein said analysis comprises identifying one or more features of said original image having one of a known or predicted orientation and determining the angle of rotation of said identified feature with respect to said known or predicted orientation.

11. A method of correcting tilt, as claimed in claim 1 wherein the crop boundary can be in one of portrait and landscape format irrespective of the notional configuration of the original image.

12. A method as claimed in claim 1, wherein the cropped image is rectangular, with an aspect ratio automatically selected by an image processor based on image content.

13. A method of correcting tilt, wherein the method comprises:

a) receiving an electronic image;

b) rotating the image to correct for tilt;

c) automatically analyzing different image features within the image in accordance with at least one compositional rule, wherein the rule is applied based on subject matter of the original image, in order to identify compositionally significant features within the image and to generate at least one crop of the image; and

d) outputting the at least one cropped image;

wherein the cropped image is not merely concentric with the original electronic image.

14. A method as claimed in claim 13, wherein the image is analyzed to identify the colors within the image and regions having a color which occupies a relatively small portion of the image are determined to be salient.

15. A method as claimed in claim 13, wherein one of the compositional rules includes ensuring that objects occurring in a central region of the image are included in the at least one cropped image.

16. A method as claimed in claim 13, wherein one of the compositional rules includes performing a search for faces and including them within the crop.

17. A method as claimed in claim 13, wherein the tilt of the image is determined by a tilt sensor at the time of image capture.

18. A method as claimed in claim 13, wherein the tilt of the image is determined by analysis of features contained within the image.

19. An image processing system for correcting tilt of an electronic image, comprising:

an image input arranged to receive an original rotated image;

an angle of rotation input arranged to receive a signal indicative of an angle of rotation associated with said original image;

an analyzer configured to analyze different image features within the original image to determine a subject matter of the original image; and

an image processor connected to said analyzer, said image input and said angle of rotation input and arranged to rotate said original image to correct for said rotation and to crop the resulting rotated image in accordance with a predefined set of image composition rules, wherein the rules are based on the subject matter of the original image as determined by said analyzer, such that the crop boundary is contained within the original image.

20. An image processing system as claimed in claim 19, wherein the image composition rules comprise:

identifying at least one region of interest within the image;

selecting at least one of the identified regions to be included within the cropped image; and

generating at least one crop boundary to include at least one of the identified regions.

21. An image processing system as claimed in claim 19, wherein the at least one cropped image is not concentric with the original rotated image.

22. An image processing system as claimed in claim 19, wherein the at least one crop boundary is smaller than the largest tilt corrected crop that can be contained within the original rotated image.

23. An image processing system as claimed in claim 19, wherein the system includes a sensor for detecting the orientation of an image capture array.

24. A camera including an image processing system as claimed in claim 19.