Method and apparatus for multifocus digital image restoration using image integration technology

• US 20070286517A1
• Filed: 12/15/2006
• Published: 12/13/2007
• Est. Priority Date: 06/13/2006
• Status: Active Grant

• Current Assignee: Chung-Ang University Inudstry-Academy Cooperation Foundation
• Original Assignee: Industry-University Cooperation Foundation Hanyang University
• Inventors: Paik, Joon-Ki, Yoo, Yoon-Jong, Vivek, Maik, Shin, Jeong-Ho
• Granted Patent: US 7,881,551 B2
• Time in Patent Office: Days
• US Class Current: 382/255
• CPC Class Codes: G06T 5/73 Deblurring; Sharpening
• CPC Class Codes: H04N 23/951 by using two or more images to influence resolution, frame rate or aspect ratio

Abstract

Methods and apparatus are disclosed for digital image restoration. The method including the steps of receiving an external deteriorated image, and generating N restoration images by using a preset PSF and an image restoration filter without computing a PSF of the input deteriorated image, dividing the restoration images into window parts of a predetermined window size, computing absolute values of differences among peripheral pixels within the divided window parts, selecting only a high frequency part, and analyzing focus deterioration; determining weights of the N restoration images, selecting window parts having high weights, and integrating the selected window parts into one image and comparing a degree of restoration of the integrated image with a preset degree of restoration, and outputting the integrated image as a resultant image when the degree of restoration of the integrated image is higher than the preset degree of restoration.

Claims

1. A multifocus digital image restoration method using an image integration technology, the method comprising the steps of:

receiving a deteriorated image, and generating N restoration images by using a preset Point Spread Function (PSF) and an image restoration filter without computing a PSF of the input deteriorated image;

dividing the N restoration images into window parts of a predetermined window size,

computing absolute values of differences among peripheral pixels within the divided window parts,

selecting only a high frequency part within the window part, and analyzing focus deterioration;

determining weights of the N restoration images by using the analyzed focus deterioration;

selecting window parts having high weights based on the determined weights;

integrating the selected window parts into one image;

comparing a degree of restoration of the integrated image with a preset degree of restoration, and

outputting the integrated image as a resultant image when the degree of restoration of the integrated image is higher than the preset degree of restoration.

2. The method as claimed in claim 1, wherein the N restoration images are generated by repetitively executing a regularization restoration formula by using one fixed PSF.

3. The method as claimed in claim 2, wherein the degree of restoration of the restoration image becomes increasingly as the regularization restoration formula is repetitively executed.

4. The method as claimed in claim 1, wherein the absolute values are determined and summed up to determine one representative value, and only a representative value of more than a predetermined size is recorded, wherein the absolute values as determined as:

M_k(i,j)=|2I(i,j)−

I(i−

1,j)−

I(i+1,j)|+|2I(i,j)−

I(i,j−

1)−

I(i,j+1)|,

wherein M(i,j) denotes the absolute values among the peripheral pixels within the divided window parts.

5. The method as claimed in claim 1, wherein the weights of the N restoration images are determined according to difference values with the peripheral pixels, which are computed and measured from the analyzed focus deterioration.

6. The method as claimed in claim 1, wherein the degree of restoration of the integrated image is compared with the preset degree of restoration, said preset degree of restoration is selected from the group consisting of:

Signal to Noise Ratio (SNR) value and a peak Signal to Noise Rate (PSNR) value.

7. A multifocus digital image restoration method using an image integration technology, the method comprising the steps of:

receiving a deteriorated image,

dividing the input deteriorated image according to preset Point Spread Functions (PSFs) having different degrees of restoration depending on a focal distance without computing a PSF of the input deteriorated image,

estimating PSFs in each divided part, generating N restoration images, and sequentially transferring the N restoration images one by one according to the focal distance;

determining whether a previously stored restoration image or an integrated image exists,

analyzing focus deterioration of the input restoration image and the previously stored image when it is determined that the previously stored image exists;

comparing the analyzed focus deterioration and determining weights of the input restoration image and the previously stored image;

integrating the input restoration image and the previously stored image based on the determined weights;

comparing a degree of restoration of the integrated image with a preset degree of restoration, and

outputting the integrated image as a resultant image when the degree of restoration of the integrated image is higher than the preset degree of restoration.

8. The method as claimed in claim 7, further comprising a step of:

when no previously stored restoration image or integrated image is determined, storing the currently input restoration image in a storage unit, and

sequentially receiving a subsequent restoration image according to the focal distance.

9. The method as claimed in claim 7, wherein, when the degree of restoration of the integrated image is lower than the preset degree of restoration, the integrated image is stored in a storage unit, a subsequent restoration image is sequentially received according to the focal distance, and the step of integrating the images is repeated.

10. The method as claimed in claim 7, wherein, when the focus deterioration of the restoration image and the previously stored image is analyzed, the input restoration image and the previously stored image are divided into window parts of a predetermined window size, respectively, absolute values of differences among peripheral pixels within the divided window parts are computed, and only a high frequency part within the window part is selected.

11. A digital multifocus restoration apparatus using an image integration technology, the apparatus comprising:

an image sensor unit for;

receiving a deteriorated image,

converting the input deteriorated image to an image on a frequency domain, and

outputting the converted deteriorated image;

an image restoration filter unit for;

receiving the converted deteriorated image from the image sensor unit,

generating N restoration images by using a preset Point Spread Function (PSF) and an image restoration filter without computing a PSF of the input deteriorated image, and

outputting the N restoration images; and

an image integration unit for;

receiving the N restoration images from the image restoration filter unit,

analyzing focus deterioration of the N restoration images,

determining weights of the N restoration images by using the analyzed focus deterioration,

integrating the images into one image based on the determined weights, and

outputting the integrated image as a resultant image when a degree of restoration of the integrated image is higher than a preset degree of restoration.

12. The apparatus as claimed in claim 11, wherein the image integration unit comprises:

a focus deterioration analyzer for;

receiving the N restoration images from the image restoration filter unit,

dividing the N restoration images into window parts of a predetermined window size,

computing absolute values of differences among peripheral pixels within the divided window parts,

selecting only a high frequency part within the window part, and analyzing focus deterioration;

a weight determiner for;

determining weights of the N restoration images by using the focus deterioration analyzed by the focus deterioration analyzer; and

a comparator for;

selecting window parts having high weights based on the weights determined by the weight determiner,

integrating the selected window parts into one image,

comparing a degree of restoration of the integrated image with a preset degree of restoration, and

outputting the integrated image as a resultant image when the degree of restoration of the integrated image is higher than the preset degree of restoration.

13. The apparatus as claimed in claim 11, wherein the input deteriorated image is converted to the image on the frequency domain in order to improve a filtering speed.

14. The apparatus as claimed in claim 11, wherein the preset PSF includes a PSF calculated in advance in an optical system design step, a PSF detected through an actual experiment, and a PSF estimated in a form of a Gauss function.

15. A multifocus digital image restoration apparatus comprising:

a processor in communication with a memory, the memory containing computer code providing instruction to the processor for executing the steps of;

receiving a deteriorated image,

generating N restoration images by using a preset Point Spread Function (PSF) and an image restoration filter without computing a PSF of the input deteriorated image;

dividing the N restoration images into window parts of a predetermined window size,

computing absolute values of differences among peripheral pixels within the divided window parts,

selecting only a high frequency part within the window part, and analyzing focus deterioration;

determining weights of the N restoration images by using the analyzed focus deterioration;

selecting window parts having high weights based on the determined weights;

integrating the selected window parts into one image;

comparing a degree of restoration of the integrated image with a preset degree of restoration; and

outputting the integrated image as a resultant image when the degree of restoration of the integrated image is higher than the preset degree of restoration.

16. The apparatus as claimed in claim 15, wherein the N restoration images are generated by repetitively executing a regularization restoration formula by using one fixed PSF.

17. The apparatus as claimed in claim 16, wherein the degree of restoration of the restoration image becomes increasingly as the regularization restoration formula is repetitively executed.

18. The apparatus as claimed in claim 15, wherein the absolute values are determined and summed up to determine one representative value, and only a representative value of more than a predetermined size is recorded, wherein the absolute values as determined as:

M_k(i,j)=|2I(i,j)−

I(i−

1,j)−

I(i+1,j)|+|2I(i,j−

1)−

I(i,j+1)|,

wherein M(i,j) denotes the absolute values among the peripheral pixels within the divided window parts.

19. The apparatus as claimed in claim 15, wherein the weights of the N restoration images are determined according to difference values with the peripheral pixels, which are computed and measured from the analyzed focus deterioration.

20. The apparatus as claimed in claim 15, wherein the degree of restoration of the integrated image is compared with the preset degree of restoration, said preset degree of restoration is selected from the group consisting of:

Signal to Noise Ratio (SNR) value and a peak Signal to Noise Rate (PSNR) value.