Method of processing an image to form an image pyramid

1. A method of processing a digital image with a computer having a digital image processing function to form an image pyramid having a first to a Nth number of pyramid image levels from a luminance image obtained from the digital image, the Nth number corresponding to a predetermined number of pyramid image levels or a number reached when a predetermined condition is met, the method comprising the steps of:

a) generating a first level base image comprising pixel values at pixel locations arranged in rows and columns based on an initial level base image, which is the luminance image obtained from the digital image, the first level base image having a smaller resolution than the initial level base image;

b) generating a next level base image by:

determining sample locations for the next level base image in the pyramid based solely on the previous level base image without adding a border around the previous level base image, such that the sample locations are arranged in a regular pattern and the sample locations both span an entire range of and beyond the pixel locations of the previous level base image but with the next level base image having a smaller resolution than the previous level base level image;

determining the pixel values of the next level base image by interpolating the pixel values of the previous level base image using an interpolation filter at the determined sample locations, wherein the interpolation filter is configured to interpolate between pixel values nearby the determined sample locations, and wherein mirroring is used for sample locations within the determined sample locations (1) whose pixel values are to be interpolated and (2) whose locations are outside of the range of the pixel locations of the previous level base image;

c) generating a subsequent level base image to an Nth level base image by repeating step b) until the Nth number of pyramid image levels are generated.

2. The method claimed in claim 1 , further comprising wherein the pyramid image includes residual images, which are formed by the steps of:

d) generating a first level residual image having the same resolution as the initial level base image by interpolating the first level base image to have the same resolution as the initial level base image, and calculating the difference between the initial level base image and the interpolated first level base image;

e) generating a next level residual image having the same resolution as the previous level base image by interpolating the next level base image to have the same resolution as the previous level base image, and calculating the difference between the previous level base image and the interpolated next level base image; and

f) generating a subsequent level residual image to an Nth level residual image by repeating step e) until the Nth level residual image is generated.

3. The method claimed in claim 1 , wherein the determined sample locations preserve the phase of the initial base level image.

4. The method claimed in claim 1 , wherein the interpolation filter is a block average filter.

5. The method claimed in claim 1 , wherein interpolation filter employs reflected pixel values at the edges of the digital image.

6. A method of processing a digital image to form an enhanced luminance image from a luminance image obtained from the digital image, with a computer having a digital image processing function, the method comprising the steps of:

a) generating an image pyramid, having a first to a Nth number of pyramid image levels from the luminance image obtained from the digital image, the Nth number corresponding to a predetermined number of pyramid image levels or a number reached when a predetermined condition is met, by:

i) generating a first level base image comprising pixel values at pixel locations arranged in rows and columns based on an initial level base image, which is the luminance image, the first level base image having a smaller resolution than the initial level base image;

ii) generating a first level residual image having the same resolution as the initial level base image by interpolating the first level base image to have the same resolution as the initial level base image, and calculating the difference between the initial level base image and the interpolated first level base image;

iii) forming a next level base image by:

determining sample locations for the next base level image in the pyramid based solely on the previous level base image without adding a border wound the previous level base image, such that the sample locations are arranged in a regular pattern and the sample locations both span an entire range of and beyond the pixel locations of the previous level base image but with the next level base image having a smaller resolution than the previous level base image;

determining the pixel values of the next level base image by interpolating the pixel values of the previous level base image using an interpolation filter at the determined sample locations, wherein the interpolation filter is configured to interpolate between pixel values nearby the determined sample locations, and wherein mirroring is used for sample locations within the determined sample locations (1) whose pixel values are to be interpolated and (2) whose locations are outside of the range of the pixel locations of the previous level base image;

iv) generating a next level residual image having the same resolution as the previous level base image by interpolating the next level base image to have the same resolution as the previous level base image and calculating the difference between the previous level base image and the interpolated next level base image; and

v) generating a subsequent level base image and a subsequent level residual image, and higher by repeating steps iii) and iv) until Nth level base and residual images are generated; and

b) generating a pedestal image from the image pyramid obtained in step a) above;

c) applying a tone scale function to the pedestal image generated in step b) to produce a tone scale adjusted pedestal image;

d) generating a texture image from the luminance image and the generated pedestal image;

e) adding the texture image obtained in step d) to the tone scale adjusted pedestal image produced in step c) to produce an enhanced luminance image.

7. The method claimed in claim 6 , wherein the step of generating the pedestal image comprises the steps of:

modifying the generated residual images to reduce the magnitude thereof in regions corresponding to non-edge regions of the initial level base image to produce modified residual images;

generating the pedestal image by reconstructing the digital image using the modified residual images.

8. The method claimed in claim 6 , wherein the sample locations preserve the phase of the initial level base image.

9. The method claimed in claim 6 , wherein the interpolation filter employs reflected pixel values at the edges of the digital image.

10. An apparatus for processing a digital image to form an image pyramid having a first to a Nth number of pyramid image levels from a luminance image obtained from the digital image, the apparatus comprising a computer readable storage medium having stored thereon a computer program product configured to:

a) generate a first level base image comprising pixel values at pixel locations arranged in rows and columns based on an initial level base image, which is the luminance image, the first level base image having a smaller resolution than the initial level base image;

b) generate a second level base image by:

i) determining sample locations for the second level base image in the pyramid based solely on the base level image without adding a border around the first level base image, such that the sample locations are arranged in a regular pattern and the sample locations both span an entire range of and beyond the pixel locations of the first level base image but with the second level base image having a smaller resolution than the first level base level image;

ii) determining the pixel values of the second level base image by interpolating the pixel values of the first level base image using an interpolation filter at the determined sample locations, wherein the interpolation filter is configured to interpolate between pixel values nearby the determined sample locations, and wherein mirroring is used for sample locations within the determined sample locations (1) whose pixel values are to be interpolated and (2) whose locations outside of the range of the pixel locations of the first level base image; and

c) generate a subsequent base level image to an Nth base level image.

11. The apparatus claimed in claim 10 , wherein the pyramid image includes residual images, the computer program product further being configured to:

generate a first level residual image having the same resolution as the initial level base image by interpolating a the first level base image to have the same resolution as the initial level base image, and calculating the difference between the initial level base image and the interpolated first level base image;

generate a second level residual image having the same resolution as the first level base image by interpolating the second level base image to have the same resolution as the first level base image, and calculating the difference between the first level base image and the interpolated second level base image; and

generate a subsequent level residual image to a Nth level residual image.

12. The apparatus claimed in claim 10 , wherein the determined sample locations preserve the phase of the initial level base image.

13. The apparatus claimed in claim 10 , wherein the interpolation filter is a block average filter

14. The apparatus claimed in claim 10 , wherein the interpolation filter employs reflected pixel values at the edges of the digital image.

15. An apparatus for processing a digital image to form an enhanced luminance image from a luminance image obtained from the digital image, the apparatus comprising a computer readable storage medium having stored thereon a computer program product configured to:

a) generate an image pyramid having a first to a Nth number of pyramid image levels from the luminance image obtained from the digital image, the Nth number corresponding to a predetermined number of pyramid image levels or a number reached when a predetermined condition is met, the image pyramid being generated by;

i)generating a first level base image comprising pixel values at pixel locations arranged in rows and columns based on an initial level base image, which is the luminance image, the first level base image having a smaller resolution than the initial level base image;

ii) generating a first level residual image having the same resolution as the initial level base image by interpolating a the first level base image to have the same resolution as the initial level base image, and calculating the difference between the initial level base image and the interpolated first level base image;

iii) generating a second level base image by:

determining sample locations for the second level base image in the pyramid based solely on the first level base image without adding a border around the first level base image, such that the sample locations are arranged in a regular pattern and the sample locations both span an entire range of and beyond the pixel locations of the first level base image but with the second level base image having a smaller resolution than the first level base level image; and

determining the pixel values of the second level base image by interpolating the pixel values of the first level base image using an interpolation filter at the determined sample locations, wherein the interpolation filter is configured to interpolate between pixel values nearby the determined sample locations, and wherein mirroring is used for sample locations within the determined sample locations (1) whose pixel values are to be interpolated and (2) whose locations are outside of the range of the pixel locations of the first level base image;

iv) generating a second level residual image having the same resolution as the first level base image by interpolating the second level base image to have the same resolution as the first level base image, and calculating the difference between the first level base image and the interpolated second level base image;

v) generating a subsequent level base image to a Nth level base image; and

vi) generating a subsequent level residual image to a Nth level residual image;

b) generate a pedestal image from the generated image pyramid;

c) apply a tone scale function to the generated pedestal image to produce a tone scale adjusted pedestal image;

d) generate a texture image from the luminance image and the generated pedestal image; and

e) add the generated texture image to the tone scale adjusted pedestal image to produce an enhanced luminance image.

16. The apparatus claimed in claim 15 , wherein the process of generating the pedestal image comprises:

modifying the generated residual images to reduce the magnitude thereof in regions corresponding to non-edge regions of the initial level base image to produce modified residual images; and

generating the pedestal image by reconstructing the digital image using the modified residual images.

17. The apparatus claimed in claim 15 , wherein the determined sample locations preserve the phase of the initial level base image.

18. The apparatus claimed in claim 15 , wherein interpolation filter employs reflected pixel values at the edges of the digital image.

19. A computer readable medium having stored thereon a computer program product for performing the method of claim 1 .

20. A computer readable medium having stored thereon a computer program product for performing the method of claim 6 .