Saliency based capture or image processing

1. A device for image capture, the device comprising:

a memory; and

one or more processors coupled to the memory and configured to:

receive, during a preview mode or a recording, a first image generated with a lens of a camera at a first lens position;

generate a first saliency map indicative of relative saliency of different regions within the first image, wherein the relative saliency of the different regions is indicative of a likelihood of attracting viewer gaze;

generate one or more additional images based on manipulating pixels in the first image;

generate one or more additional saliency maps indicative of relative saliency of different regions within the one or more additional images;

determine, during the preview mode or the recording, an autofocus setting that defines a second lens position for the lens based on a comparison of at least the first saliency map to at least one of the additional saliency maps;

determine that the second lens position and the first lens position is different; and

adjust a lens position of the lens to the second lens position.

2. The device of claim 1 , wherein the one or more additional images comprise a first set of one or more additional images, wherein the one or more additional saliency maps comprise a first set of one or more additional saliency maps, wherein the autofocus setting comprises a first autofocus setting, wherein the preview mode or the recording comprises a first preview mode or recording, and wherein the one or more processors are configured to:

receive, during a second preview mode or recording, a second image generated with the lens of the camera at a third lens position;

generate a second saliency map indicative of relative saliency of different regions within the second image;

generate a second set of one or more additional images based on manipulating pixels in the second image;

generate a second set of one or more additional saliency maps indicative of relative saliency of different regions within the second set of one or more additional images;

determine, during the second preview mode or recording, a second autofocus setting that defines a fourth lens position for the lens based on a comparison of at least the second saliency map to at least one of the second set of one or more additional saliency maps;

determine that the fourth lens position and the third lens position is the same; and

avoid adjustment of a lens position of the lens.

3. The device of claim 1 , wherein generating the one or more additional images based on manipulating pixels in the first image comprises generating the one or more additional images based on depth of image content in the first image.

4. The device of claim 1 , wherein generating the one or more additional images comprises generating the one or more additional images based on manipulating pixels of objects in a foreground of the first image.

5. The device of claim 1 , wherein comparing the first saliency map to the one or more additional saliency maps comprises one or more of performing a cross correlation, a sum of absolute difference process, or a mean square error process.

6. The device of claim 1 , wherein determining the autofocus setting that defines the second lens position for the lens based on the comparison of at least the first saliency map to at least one of the additional saliency maps comprises:

determining that the first saliency map and the one or more additional saliency maps are substantially the same; and

determining the autofocus setting based on regions having relative saliency in the first saliency map and the one or more additional saliency maps.

7. The device of claim 1 , wherein determining the autofocus setting that defines the second lens position for the lens based on the comparison of at least the first saliency map to at least one of the additional saliency maps comprises:

determining that the first saliency map and the one or more additional saliency maps are not substantially the same;

determining foreground areas in the first image; and

determining an autofocus setting based on the foreground areas.

8. The device of claim 1 , wherein the one or more additional images comprises a first additional image and a second additional image, and wherein generating one or more additional images based on manipulating pixels in the first image comprises:

manipulating pixels of the first image to generate the first additional image; and

manipulating pixels of the first additional image to generate the second additional image.

9. The device of claim 1 , wherein generating the one or more additional images comprises inpainting the first image to generate the one or more additional images.

10. The device of claim 1 , wherein generating the first saliency map comprises:

downscaling the first image to generate a N×M sized downscaled image; and

generating the first saliency map based on the downscaled image,

wherein a size of the first saliency map is X×Y, and

wherein at least one of X is less than N or Y is less than M.

11. The device of claim 1 , wherein determining the autofocus setting that defines the second lens position for the lens comprises:

determining a most salient depth based on the comparison of at least the first saliency map to at least one of the additional saliency maps; and

determining the autofocus setting that defines the second lens position for the lens based on the determined most salient depth.

12. The device of claim 1 , wherein the device is one or more of a digital camera, a digital video camcorder, or a camera-equipped wireless communication device handset.

13. The device of claim 1 , wherein determining the camera setting based on the comparison of at least the first saliency map to at least one of the additional saliency maps comprises:

determining that a difference between the first saliency map and the one or more additional saliency maps is less than a threshold; and

determining the autofocus setting based on regions having relative saliency in the first saliency map and the one or more additional saliency maps.

14. The device of claim 13 ,

wherein the one or more additional images comprise a first set of one or more additional images,

wherein the one or more additional saliency maps comprise a first set of one or more additional saliency maps,

wherein the camera setting is a first camera setting,

wherein the autofocus setting a first autofocus setting, and

wherein the one or more processors are configured to:

receive a second image;

generate a second saliency map indicative of relative saliency of different regions within the second image;

generate a second set of one or more additional images based on manipulating pixels in the second image;

generate a second set of one or more additional saliency maps indicative of relative saliency of different regions within the second set of one or more additional images; and

determine that a difference between the second saliency map and the second set of one or more additional saliency maps is greater than the threshold; and

determine a second autofocus setting based on foreground areas in the second image.

15. A method for image capture, the method comprising:

receiving, during a preview mode or a recording, a first image;

generating a first saliency map indicative of relative saliency of different regions within the first image, wherein the relative saliency of the different regions is indicative of a likelihood of attracting viewer gaze;

generating one or more additional images based on manipulating pixels in the first image;

generating one or more additional saliency maps indicative of relative saliency of different regions within the one or more additional images; and

determining, during the preview mode or the recording, a camera setting based on a comparison of at least the first saliency map to at least one of the additional saliency maps wherein determining the camera setting based on the comparison of at least the first saliency map to at least one of the additional saliency maps comprises:

determining that the first saliency map and the one or more additional saliency maps are substantially the same; and

determining an autofocus setting based on regions having relative saliency in the first saliency map and the one or more additional saliency maps.

16. The method of claim 15 , wherein the first image is generated with a lens of a camera at a first lens position, and wherein determining the autofocus setting comprises determining the autofocus setting that defines a second lens position for the lens.

17. The method of claim 16 , further comprising:

determining that the second lens position and the first lens position is the same; and

avoiding adjustment of a lens position of the lens.

18. The method of claim 16 , further comprising:

determining that the second lens position and the first lens position is different; and

adjusting a lens position of the lens to the second lens position.

19. The method of claim 15 , wherein generating the one or more additional images based on manipulating pixels in the first image comprises generating the one or more additional images based on depth of image content in the first image.

20. The method of claim 15 , wherein generating the one or more additional images comprises generating the one or more additional images based on manipulating pixels of objects in a foreground of the first image.

21. The method of claim 15 , wherein comparing the first saliency map to the one or more additional saliency maps comprises one or more of performing a cross correlation, a sum of absolute difference process, or a mean square error process.

22. The method of claim 15 , wherein the one or more additional images comprise a first set of one or more additional images, wherein the one or more additional saliency maps comprise a first set of one or more additional saliency maps, wherein the autofocus setting comprises a first autofocus setting, wherein the preview mode or the recording comprises a first preview mode or recording, wherein the camera setting comprises a first camera setting, the method further comprising:

receiving, during a second preview mode or recording, a second image;

generating a second saliency map indicative of relative saliency of different regions within the second image;

generating a second set of one or more additional images based on manipulating pixels in the second image;

generating a second set of one or more additional saliency maps indicative of relative saliency of different regions within the second set of one or more additional images;

determining, during the second preview mode or recording, a second autofocus setting that defines a fourth lens position for the lens based on a comparison of at least the second saliency map to at least one of the second set of one or more additional saliency maps;

determining a second camera setting based on the comparison of at least the second saliency map to at least one of the second set of one or more additional saliency maps, wherein determining the second camera setting based on the comparison of at least the second saliency map to at least one of the second set of one or more additional saliency maps comprises:

determining that the second saliency map and the second set of one or more additional saliency maps are not substantially the same;

determining foreground areas in the second image; and

determining a second autofocus setting based on the foreground areas.

23. The method of claim 15 , wherein generating one or more additional images comprises inpainting the first image to generate the one or more additional images.

24. The method of claim 15 , wherein determining the camera setting comprises:

determining a most salient depth based on the comparison of at least the first saliency map to at least one of the additional saliency maps; and

determining the autofocus setting comprises determining the autofocus setting based on the determined most salient depth.

25. The method of claim 15 ,

wherein the one or more additional images comprise a first set of one or more additional images,

wherein the one or more additional saliency maps comprise a first set of one or more additional saliency maps,

wherein the camera setting is a first camera setting, and

wherein the autofocus setting a first autofocus setting,

the method further comprising:

receiving a second image;

generating a second saliency map indicative of relative saliency of different regions within the second image;

generating a second set of one or more additional images based on manipulating pixels in the second image;

generating a second set of one or more additional saliency maps indicative of relative saliency of different regions within the second set of one or more additional images; and

determining that a difference between the second saliency map and the second set of one or more additional saliency maps is greater than a threshold; and

determining a second autofocus setting based on foreground areas in the second image.

26. A computer-readable storage medium storing instructions thereon that when executed cause one or more processors to:

receive, during a preview mode or a recording, a first image generated with a lens of a camera at a first lens position;

generate a first saliency map indicative of relative saliency of different regions within the first image, wherein the relative saliency of the different regions is indicative of a likelihood of attracting viewer gaze;

generate one or more additional images based on manipulating pixels in the first image;

generate one or more additional saliency maps indicative of relative saliency of different regions within the one or more additional images;

determine, during the preview mode or the recording, a camera setting an autofocus setting that defines a second lens position for the lens based on a comparison of at least the first saliency map to at least one of the additional saliency maps;

determine that the second lens position and the first lens position is different; and

adjust the lens position of the lens to the second lens position.

27. A device for image capture, the device comprising:

a memory; and

one or more processors coupled to the memory and configured to:

receive, during a preview mode or a recording, a first image;

generate a first saliency map indicative of relative saliency of different regions within the first image, wherein the relative saliency of the different regions is indicative of a likelihood of attracting viewer gaze;

simulate different exposures on the first image based on changing tone of the first image to generate one or more additional images;

generate one or more additional saliency maps, indicative of relative saliency of different regions within the one or more additional images, within the one or more additional images;

generate a plurality of metering maps based on the first saliency map and the one or more additional saliency maps;

determine an updated metering map based on the plurality of metering maps; and

determine, during the preview mode or the recording, an autoexposure setting based on the updated metering map.

28. A method for image capture, the method comprising:

receiving, during a preview mode or a recording, a first image;

generating a first saliency map indicative of relative saliency of different regions within the first image, wherein the relative saliency of the different regions is indicative of a likelihood of attracting viewer gaze;

simulating different exposures on the first image based on changing tone of the first image to generate one or more additional images;

generating one or more additional saliency maps, indicative of relative saliency of different regions within the one or more additional images, within the one or more additional images;

generating a plurality of metering maps based on the first saliency map and the one or more additional saliency maps;

determining an updated metering map based on the plurality of metering maps; and

determining, during the preview mode or the recording, an autoexposure setting based on the updated metering map.