Method and apparatus for improved image interpolation

• US 6,832,009 B1
• Filed: 09/24/1999
• Issued: 12/14/2004
• Est. Priority Date: 09/24/1999
• Status: Expired due to Term

• Current Assignee: Qualcomm technologies international, ltd. (Qualcomm, Inc.)
• Original Assignee: Zoran Corporation (Qualcomm, Inc.)
• Inventors: Sutskover, Ilan, Shezaf, Nira, Abramov-Segal, Hagit, Bar-Sella, Ran
• Primary Examiner(s): Patel, Kanji
• Application Number:
  US09/405,714
• Time in Patent Office: 1,908 Days
• Field of Search: 382/300, 382/284, 382/276, 382/293, 382/298, 382/299, 348/240.99, 348/365, 348/406.1, 345/606, 345/608, 345/609, 345/610, 345/611, 345/617, 358/525, 358/528
• US Class Current: 382/300
• CPC Class Codes: G06T 3/403 Edge-driven scaling; Edge-based scaling

Abstract

A method and apparatus for image interpolation provides a simple method that can up-scale an image in various scale ratios, including fractional scaling ratios, and can yield a sharp image with reduced unpleasant visual artifacts. An embodiment uses a power weight function that is gradient dependent and uses a function that has a linear dependence on the distance to compute interpolated pixels. A further embodiment uses a power that can be adapted according to the local contrast of edges.

Claims

1. A method for determining an interpolated pixel value comprising:

determining a plurality of gradient values for a plurality of original pixels near said interpolated pixel;

determining a plurality of distance values for a plurality of original pixels near said interpolated pixel;

combining a plurality of original pixel values, at least some of said original pixel values adjusted by said gradient values and said distance values; and

normalizing the result of said combining to produce a value of said interpolated pixel.

2. The method of claim 1 wherein said gradient value is incorporated into a power function prior to said combining.

3. The method of claim 2 wherein said power function is adaptive according to local contrast of edges.

4. The method of claim 1 wherein said distance value has a linear dependence on the distance between said pixel and surrounding prior pixels.

5. The method of claim 1 wherein said gradient values are computed using a Sobel operation.

6. The method of claim 1 wherein said gradient values are computed using a Prewitt operation.

7. The method of claim 1 wherein said gradient values are computed using a Roberts operation.

8. The method of claim 1 wherein said method is performed by an appropriately configured digital processing device.

34. The method of claim 1, additionally comprising adding the value of the interpolated pixel to values of the original pixels, thereby to increase the number of values of pixels.

9. A method carried out by electronic circuitry for determining a pixel value p(c) of calculated pixel c from the values p(a_i) of surrounding original pixels a_{i(l to n}) according to an equation of the form:

$p<br><br>\left(c\right)=\frac{\underset{i=1<br><br>\mathrm{tol}}{\sum <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>W<br><br>\left(a_i\right)<br><br>p<br><br>\left(a_i\right)}{\underset{i=1<br><br>\mathrm{tol}}{\sum <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>W<br><br>\left(a_i\right)}$

where W(a_i) is a function with linear dependence on the distance and dependent on a local gradient, wherein the calculation according to said equation is performed by the electronic circuitry.

10. The method of claim 9 where 1=4.

11. The method of claim 9 where W(a_i)=D(a_i)G′

(a_i).

12. The method of claim 11 wherein D(a_i) is given by an expression of the form:

D(a_i)=(1−

Δ

x_i)(1−

Δ

y_i).

13. The method of claim 11 wherein G′

(a_i) is given by an expression of the form;

G′

(a_i)=(−

μ

G(a_i)+1)″

.

14. The method of claim 13 wherein G(a_i) is a gradient value computing from a plurality of surrounding pixels.

15. The method of claim 13 wherein G(a_i) is a gradient value computing using a convolution kernel from a plurality of surrounding original pixels.

16. The method of claim 13 wherein G(a_i) is a gradient value computing using a Sobel operator.

17. The method of claim 13 wherein G(a_i) is a gradient value computing using a Prewitt operator.

18. The method of claim 13 wherein G(a_i) is a gradient value computing using a Roberts operator.

19. The method of claim 13 wherein μ

is a positive constant value close to and lower than 1.

20. The method of claim 13 wherein n is a constant value selected for a particular image.

21. The method of claim 13 wherein n is selected based on the average gradients of edges in said image.

22. The method of claim 13 wherein n is a value that is selected for a pixel based on the local contrast at said pixel.

23. The method of claim 22 wherein n is a value that is selected for a pixel based on the local contrast at said pixel according to the formula:

$n=\frac{\mathrm{base\_n}}{\max (G<br><br>\left(a_1\right),\mathrm{\dots <br><br>}<br><br>\text{\hspace{1em}<br><br>},G<br><br>\left(a_k\right)<br><br>\left(\mathrm{\alpha <br><br>}\right)}$

24. The method of claim 23 wherein k=4.

25. The method of claim 23 wherein α

is a constant between 0 and 1 and is chosen so that base_n/Alpha is the maximum power and where base_n is a minimum power chosen so that the power fits the highest edge contrast that can be in the image.

26. The method of claim 23 wherein said method is performed by an appropriately configured digital processing device.

27. A method for determining a pixel value for an interpolating pixel in an image comprising:

determining a first sum of a plurality of values of a plurality of surrounding pixels, said plurality of values modified by at least two weighting factors;

dividing said first sum by a second sum, said second sum comprising said weighting factors wherein one of said weighting factors is a power weight function that is gradient dependent; and

developing an electronic image signal which incorporates the interpolated image pixel value therein.

28. The method of claim 27 wherein one of said weighting factors has a linear dependence on the distance.

29. The method of claim 27 wherein one of said weighting factors is a power weight function that is adaptive according to the local contrast of edges.

30. A digital image capture device comprising:

image capture sensors for capturing an image at a first pixel density;

image storage for holding a portion up to all of said captured image;

logic components for increasing pixel density and determining the value of an interpolated pixel by;

determining a plurality of gradient values for a plurality of original pixels near said interpolated pixel;

determining a plurality of distance values for a plurality of original pixels near said interpolated pixel; and

combining a plurality of original pixel values, at least some of said original pixel values adjusted by said gradient values and said distance values to produce a value of said interpolated pixel; and

an interface for outputting said image at a second pixel density.

32. The device according to claim 30, wherein the interface for outputting said image causes the second pixel density of said image to be increased.

31. An image display device capable of displaying an image with an altered pixel density comprising:

an interface for accepting an image at a first pixel density;

image storage for holding a portion of said image up to all of said image;

logic components for increasing pixel density and determining the value of an interpolated pixel by;

determining a plurality of gradient values for a plurality of original pixels near said interpolated pixel;

determining a plurality of distance values for a plurality of original pixels near said interpolated pixel; and

combining a plurality of original pixel values, at least some of said original pixel values adjusted by said gradient values and said distance values to produce a value of said interpolated pixel; and

a display for displaying said image at a second pixel density.

33. The device according to claim 31, wherein the interface for outputting said image causes the second pixel density of said image to be increased.

35. A method of determining a value of a given pixel of an image from values of pixels surrounding the given pixel, comprising:

summing values of a plurality of said surrounding pixels to determine the value of the given pixel of the image, the summed values being modified by a power weighting function that is adaptive to characteristics of the image, and

developing an electronic video signal for displaying the given pixel with the determined value.