Image processing device and image processing method

Patent metadata

Field	Value
Publication number	US-9679370-B2
Application number	US-201314648830-A
Country	US
Kind code	B2
Filing date	Dec 5, 2013
Priority date	Dec 6, 2012
Publication date	Jun 13, 2017
Grant date	Jun 13, 2017

Abstract

The present invention provides an image processing device whereby the probability of outputting a restored image which accurately corresponds to an original image which is included in a low-quality input image is improved. This image processing device comprises: an image group generating means for generating, from the input image, using a dictionary which stores a plurality of patch pairs wherein a degradation patch which is a patch of a degraded image wherein a prescribed image is degraded is associated with a restoration patch which is a patch of this prescribed image, a plurality of restored image candidates including a plurality of different instances of content which have a possibility of being the original content of the input image; and an image selection presentation means for clustering the generated plurality of restored image candidates, and selecting and outputting an image candidate on the basis of the result of this clustering.

First claim

What is claimed is:

1. An image processing device comprising:
a processor configured to:
use a dictionary storing a plurality of patch pairs including a degraded patch which is a patch of a degraded image obtained by degrading a predetermined image and a restored patch which is a patch of said predetermined image in such a manner that the degraded patch and the restored patch are associated with each other, and generate a plurality of restored image candidates including a plurality of different contents that may be an original content of an input image from said single input image;
generate a reduced-quality image from said restored image, and calculate a degree of reliability of the restored image candidate on the basis of a difference between said reduced-quality image and said input image; and
select and output said restored image candidate on the basis of said degree of reliability.

2. The image processing device according to claim 1 , wherein said processor is further configured to calculate said degree of reliability on the basis of a difference between restored patches adjacent to each other in a case where said restored image is generated.

3. The image processing device according to claim 1 , wherein said processor is further configured to calculate said degree of reliability on the basis of a degree of risk which is a ratio of a similarity degree between said degraded patches and a similarity degree between said restored patches corresponding to said degraded patches.

4. The image processing device according to claim 1 , wherein said processor is further configured to calculate said degree of reliability on the basis of a statistical deviation of said restored image candidates.

5. The image processing device according to claim 1 , wherein said processor is further configured to calculate said degree of reliability for each class of said clustered restored image candidates, and
said processor is further configured to select and output said restored image candidate for each of said classes on the basis of the degree of reliability calculated for each of said classes.

6. The image processing device according to claim 1 , wherein said processor is further configured to cut out a particular portion restored image candidate which is a particular portion of each of said restored image candidates, cluster said particular portion restored image candidates, and select and output said particular portion restored image candidate on the basis of said result of clustering.

7. The image processing device according to claim 6 , wherein said processor is further configured to calculate the degree of reliability of said particular portion restored image candidate.

8. The image processing device according to claim 1 , wherein said processor is further configured to divide each of said restored image candidates into a plurality of divided restored image candidates, cluster each of a group made up with said divided restored image candidates corresponding to a same position, select said divided restored image candidate for each of said groups on the basis of said result of clustering, and generate and output a restored image by compositing said selected divided restored image candidate.

9. The image processing device according to claim 8 , wherein said processor is further configured to calculate the degree of reliability of said divided restored image candidate.

10. The image processing device according to claim 1 , wherein said processor is further configured to generate, from said input image, said restored image candidates respectively corresponding to maximum values of said probability distribution functions of which hyperparameters are different.

11. The image processing device according to claim 1 , wherein said processor is further configured to generate, from said input image, said restored image candidates respectively corresponding to a plurality of local maximum values of probability distribution functions made by modeling a relationship of said input image, a restored image candidate corresponding to said input image, and a patch pair.

12. The image processing device according to claim 1 , wherein said processor is further configured to generate, from said input image, said restored image candidates respectively corresponding to the local maximum values of said probability distribution functions of which hyperparameters are different.

13. The image processing device according to claim 1 , wherein said processor is further configured to select a plurality of restored patches corresponding to a plurality of degraded patches of which similarity degree with respect to said input patch is equal to or more than a threshold value for each input patch that is cut out from said input image, and generate said restored image candidates by combining said selected restored patches respectively corresponding to said input patches.

14. The image processing device according to claim 1 further comprising:
a storage device that stores said patch pairs,
said processor is further configured to generate said patch pairs, and output said patch pairs to said storage device.

15. An image processing method executed by a computer, comprising:
using a dictionary storing a plurality of patch pairs including a degraded patch which is a patch of a degraded image obtained by degrading a predetermined image and a restored patch which is a patch of said predetermined image in such a manner that the degraded patch and the restored patch are associated with each other to generate a plurality of restored image candidates including a plurality of different contents that may be an original content of an input image from said single input image;
generating a reduced-quality image from said restored image;
calculating a degree of reliability of the restored image candidate on the basis of a difference between said reduced-quality image and said input image; and
selecting and outputting said restored image candidate on the basis of said result of clustering.

16. A non-transitory computer-readable recording medium recorded with an image processing program causing a computer to execute:
using a dictionary storing a plurality of patch pairs including a degraded patch which is a patch of a degraded image obtained by degrading a predetermined image and a restored patch which is a patch of said predetermined image in such a manner that the degraded patch and the restored patch are associated with each other to generate a plurality of restored image candidates including a plurality of different contents that may be an original content of an input image from said single input image;
generating a reduced-quality image from said restored image;
calculating a degree of reliability of the restored image candidate on the basis of a difference between said reduced-quality image and said input image; and
selecting and outputting said restored image candidate on the basis of said result of clustering.

17. An image processing device comprising:
image group generation means for using a dictionary storing a plurality of patch pairs including a degraded patch which is a patch of a degraded image obtained by degrading a predetermined image and a restored patch which is a patch of said predetermined image in such a manner that the degraded patch and the restored patch are associated with each other to generate a plurality of restored image candidates including a plurality of different contents that may be an original content of an input image from said single input image;
reliability degree calculation means for calculating a degr

Assignees

• Nec Corp

Inventors

• Shibata Takashi
• Iketani Akihiko
• Senda Shuji

Classifications

• G06V30/1914

  Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries · CPC title

• G06V30/164 Primary

  Noise filtering · CPC title

• G06V30/19173

  Classification techniques · CPC title

• G06T5/50 Primary

  using two or more images, e.g. averaging or subtraction · CPC title

• G06F18/2321

  using statistics or function optimisation, e.g. modelling of probability density functions · CPC title