Systems and methods for sampling an image sensor

• US 20050243183A1
• Filed: 04/30/2004
• Published: 11/03/2005
• Est. Priority Date: 04/30/2004
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Obrador, Pere
• Granted Patent: US 7,483,059 B2
• Time in Patent Office: Days
• US Class Current: 348/222.100
• CPC Class Codes: H04N 23/611 where the recognised objects include parts of the human body
• CPC Class Codes: H04N 25/443 by reading pixels from selected 2D regions of the array, e.g. for windowing or digital zooming
• CPC Class Codes: H04N 25/445 by skipping some contiguous pixels within the read portion of the array

Abstract

Systems and methods according to the present invention provide techniques which adjust the sampling scheme associated with an image sensor in order to provide high quality renditions, e.g., printed or displayed outputs. The sampling scheme can be varied on a frame by frame basis (or less frequently) to output more samples from the image sensor for regions which are predicted to contain more perceptually significant information and to output fewer samples for regions which are predicted to be less perceptually significant. Various parameters can be used as proxies for perceptual significance. For example, in images which contain people, faces can be characterized as more perceptually significant and regions which are predicted to contain faces can be sampled more densely than regions which are not predicted to contain faces. Alternatively, the spatial frequency content of data captured by an image sensor can be measured to identify regions of higher and lower spatial frequency content. Regions with higher spatial frequency content will typically contain perceptually more significant image features, e.g., edges and textures, than regions having lower spatial frequency content. Accordingly, the image sensor can be programmed to sample the former more densely.

Claims

1. A method for sampling an image sensor comprising the steps of:

evaluating a captured image;

determining a sampling allocation for said image sensor based upon said step of evaluating; and

sampling a first region of said image sensor at a first sampling density and a second region of said image sensor at a second sampling density based upon said determined sampling allocation.

2. The method of claim 1, wherein said step of evaluating further comprises the step of:

processing said captured image to determine spatial frequency content associated with pixel data captured by said image sensor.

3. The method of claim 2, wherein said step of determining a sampling allocation further comprises the step of:

assigning said first sampling density to said first region of said image sensor which has a lower spatial frequency content than that associated with said second region, wherein said first sampling density is less than said second sampling density.

4. The method of claim 1, wherein said step of evaluating further comprises the step of:

processing said captured image to determine a region containing a face in said captured image.

5. The method of claim 4, wherein said first region includes said face and said second region includes a remainder and wherein said first sampling density is greater than said second sampling density.

6. The method of claim 1, wherein said step of evaluating further comprises the step of:

evaluating said captured image to determine a plurality of regions, each of said plurality of regions having a different perceived quality value; and

wherein said step of determining further comprises the step of;

allocating more samples per unit area to those of said plurality of regions having a higher perceived quality value and allocating fewer samples per unit area to those of said plurality of regions having a lower perceived quality value.

7. The method of claim 1, wherein said step of determining said sample allocation further comprises the step of:

determining said sample allocation as a function of a bandwidth associated with a data transfer mechanism used to output samples from said image sensor and quantitative quality values output as a result of said evaluating step.

8. The method of claim 7, wherein said function is a Lagrangian optimization.

9. A computer-readable medium containing a program that performs the steps of:

evaluating a captured image;

determining a sampling allocation for said image sensor based upon said step of evaluating; and

sampling a first region of said image sensor at a first sampling density and a second region of said image sensor at a second sampling density based upon said determined sampling allocation.

10. The computer-readable medium of claim 9, wherein said step of evaluating further comprises the step of:

processing said captured image to determine spatial frequency content associated with pixel data captured by said image sensor.

11. The computer-readable medium of claim 10, wherein said step of determining a sampling allocation further comprises the step of:

assigning said first sampling density to said first region of said image sensor which has a lower spatial frequency content than that associated with said second region, wherein said first sampling density is less than said second sampling density.

12. The computer-readable medium of claim 9, wherein said step of evaluating further comprises the step of:

processing said captured image to determine a region containing a face in said captured image.

13. The computer-readable medium of claim 12, wherein said first region includes said face and said second region includes a remainder and wherein said first sampling density is greater than said second sampling density.

14. The computer-readable medium of claim 9, wherein said step of evaluating further comprises the step of:

evaluating said captured image to determine a plurality of regions, each of said plurality of regions having a different perceived quality value; and

wherein said step of determining further comprises the step of;

allocating more samples per unit area to those of said plurality of regions having a higher perceived quality value and allocating fewer samples per unit area to those of said plurality of regions having a lower perceived quality value.

15. The computer-readable medium of claim 9, wherein said step of determining said sample allocation further comprises the step of:

determining said sample allocation as a function of a bandwidth associated with a data transfer mechanism used to output samples from said image sensor and quantitative quality values output as a result of said evaluating step.

16. The computer-readable medium of claim 15, wherein said function is a Lagrangian optimization.

17. An image processing device comprising:

an image sensor for capturing an image, wherein said image sensor includes an array of image sensing pixels, wherein said image sensor is controllable to vary a sampling density for at least two regions across said array of image sensing pixels;

a processor for evaluating a captured image and for allocating a first sampling density to a first region of said image sensor and a second sampling density to a second region of said image sensor based on said evaluation of said captured image; and

a data transfer mechanism for transferring samples associated with a next image from said image sensor to said processor, wherein said samples are generated based on said first and second sampling densities.

18. The image processing device of claim 17, wherein said processor evaluates said captured image to determine spatial frequency content associated with pixel data captured by said image sensor.

19. The image processing device of claim 18, wherein processor assigns said first sampling density to said first region of said image sensor which has a lower spatial frequency content than that associated with said second region, wherein said first sampling density is less than said second sampling density.

20. The image processing device of claim 17, wherein said processor evaluates said captured image to determine a region containing a face in said captured image.

21. The image processing device of claim 20, wherein said first region includes said face and said second region includes a remainder and wherein said first sampling density is greater than said second sampling density.

22. The image processing device of claim 17, wherein said processor evaluates said captured image to determine a plurality of regions, each of said plurality of regions having a different perceived quality value and allocates more samples per unit area of said image sensor to those of said plurality of regions having a higher perceived quality value and allocates fewer samples per unit area of said image sensor to those of said plurality of regions having a lower perceived quality value.

23. The image processing device of claim 17, wherein said processor determines said sample allocation as a function of a bandwidth associated with said data transfer mechanism and quantitative quality values output as a result of said evaluating step.

24. The image processing device of claim 23, wherein said function is a Lagrangian optimization.

25. An image processing system comprising:

means for evaluating a captured image;

means for determining a sampling allocation for said image sensor based upon said step of evaluating; and

means for sampling a first region of said image sensor at a first sampling density and a second region of said image sensor at a second sampling density based upon said determined sampling allocation.