Automatically suggesting regions for blur kernel estimation

Patent metadata

Field	Value
Publication number	US-10032258-B2
Application number	US-201715452620-A
Country	US
Kind code	B2
Filing date	Mar 7, 2017
Priority date	Oct 23, 2013
Publication date	Jul 24, 2018
Grant date	Jul 24, 2018

Abstract

Embodiments of the present invention provide systems, methods, and computer storage media directed towards automatic selection of regions for blur kernel estimation. In one embodiment, a process divides a blurred image into a regions. From these regions a first region and a second region can be selected based on a number of edge orientations within the selected regions. A first blur kernel can then be estimated based on the first region and a second blur kernel can be estimated for the second region. The first and second blur kernel can then be utilized to respectively deblur a first and second portion of the image to produce a deblurred image. Other embodiments may be described and/or claimed.

First claim

The invention claimed is:

1. A computer-implemented method for processing an image comprising:
accessing the image, wherein the image includes a first blurred region and a second blurred region;
determining a positioning of a first blur kernel with respect to the first blurred region;
determining a positioning of a second blur kernel with respect to the second blurred region based on the positioning of the first blur kernel; and
deblurring the image based on the positioning of the first blur kernel and the positioning of the second blur kernel.

2. The method of claim 1 further comprising:
determining a latent image based on the image;
generating the first blur kernel based on the latent image; and
generating the second blur kernel based on the latent image.

3. The method of claim 1 further comprising:
receiving an input from a user that specifies the first blurred region and the second blurred region.

4. The method of claim 1 further comprising:
determining a region size for each of a plurality of regions of the image based on an input blur kernel size, wherein the plurality of regions includes the first blurred region and the second blurred region;
determining a number of edge orientations within each of the plurality of regions;
selecting the first blurred region for deblurring based on the number of edge orientations within the first blurred region; and
selecting the second blurred region for deblurring based on the number of edge orientations within the second blurred.

5. The method of claim 1 further comprising:
determining the positioning of the first blur kernel based on a proximity to at least one of a center of the image or a focal point of the image;
determining a shift from the positioning of the first blur kernel based on a correlation between the first blur kernel and the second blur kernel;
determining the positioning of the second blur kernel based on the shift from the positioning if the first blur kernel.

6. The method of claim 1 further comprising:
generating a deconvolution of the first blurred region based on the positioning of the first blur kernel;
generating a deconvolution of the second blurred region based on the positioning of the second blur kernel;
for an overlapping region that is included in both the first blurred region and the second blurred region, generating a blending of the deconvolution of the first blurred region and the deconvolution of the second blurred region; and
deblurring the image based on the deconvolution of the first blurred region, the deconvolution of the second blurred region, and the blending of the deconvolution of the first blurred region and the deconvolution of the second blurred region.
determining the positioning of the second blur kernel based on the shift from the positioning if the first blur kernel.

7. The method of claim 1 further comprising:
determining a first size of the first blur kernel based on an autocorrelation of the image with a derivative of the image;
determining a second size of the first blur kernel based on a latent image of the image;
determining a suggested size of the first blur kernel based on the first size and the second size of the first blur kernel; and
generating the first blur kernel based on the suggested size of the first blur kernel.

8. One or more non-transitory computer-readable media having instructions stored thereon, which, when executed by one or more processors of a computing system cause the computing system to perform actions comprising:
accessing an image that includes a first blurred region and a second blurred region;
steps for generating a first blur kernel for the first blurred region; and
steps for generating a second blur kernel for the second blurred region; and
steps for deblurring the image based on the first blur kernel and the second blur kernel.

9. The computer-readable media of claim 8 , wherein the actions further comprise:
steps for positioning the first blur kernel with respect to the first blurred region; and
steps for positioning the second blur kernel with respect to the second blurred region.

10. The computer-readable media of claim 8 , wherein the steps for deblurring the image include:
generating a deconvolution of the first blurred region based on a positioning of the first blur kernel proximate to the first blurred region;
generating a deconvolution of the second blurred region based on a positioning of the second blur kernel proximate to the second blurred region; and
generating a deblurred image based on a blending of the deconvolution of the first blurred region and the deconvolution of the second blurred region.

11. The computer-readable media of claim 8 , wherein the steps for generating the first blur kernel include:
steps for determining a first size of the first blur kernel based on an autocorrelation of the image with a derivative of the image;
steps for determining a second size of the first blur kernel based on a latent image of the image;
steps for determining a suggested size of the first blur kernel based on the first size and the second size of the first blur kernel; and
generating the first blur kernel based on the suggested size of the first blur kernel.

12. The computer-readable media of claim 11 , wherein the steps for determining the first size of the first blur kernel include:
steps for determining a derivative of the first blurred region;
steps for determining an autocorrelation between the first blurred region and the derivative of the first blurred region;
steps for applying a filter to the autocorrelation between the first blurred region and the derivative of the first blurred region;
steps for identifying one or more connected components within the autocorrelation between the first blurred region and the derivative of the first blurred region; and
determining the first size of the first blur kernel based on the identified connected components.

13. The computer-readable media of claim 11 , wherein the steps for determining the second size of the first blur kernel include:
steps for determining a latent image of the first blurred region;
steps for estimating another blur kernel based on the latent image of the first blurred region;
steps for determining a measurement of one or more connected components in the other blur kernel; and
determining the second size of the first blur kernel based on the measurement of the one or more connected components in the other blur kernel.

14. The computer-readable media of claim 8 , wherein the actions further comprise:
steps for determining a region size for each of a plurality of regions of the image based on an input blur kernel size, wherein the plurality of regions includes the first blurred region and the second blurred region;
steps for determining a number of edge orientations within each of the plurality of regions;
steps for selecting the first blurred region for deblurring based on the number of edge orientations the first blurred region; and
steps for selecting the second blurred region for deblurring based on the number of edge orientations within the second blurred region.

15. A computing system comprising:
one or more processors; and
memory, coupled with the one or more processors, having instructions stored thereon, which, when executed by the one or more processors perform actions comprising:
receiving an image;
determining a region size for each of a plurality of regions of the received image based on an input blur kernel size;
determining a number of edge orientations within each of the plurality of regions;
selecting a

Assignees

• Adobe Systems Inc
• Adobe System Incorporated

Inventors

• Cho Sunghyun
• Wang Jue
• Chien Jen-Chan
• Feng Dong

Classifications

• G06T11/10

  Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) · CPC title

• G06T2207/20192

  Edge enhancement; Edge preservation · CPC title

• G06T2207/20221

  Image fusion; Image merging · CPC title

• G06T2207/10024

  Color image · CPC title

• G06T5/20

  using local operators · CPC title