Method, apparatus and computer program for image processing including generating a different image of at least one object

Patent metadata

Field	Value
Publication number	US-11244457-B2
Application number	US-201916671214-A
Country	US
Kind code	B2
Filing date	Nov 1, 2019
Priority date	Nov 9, 2018
Publication date	Feb 8, 2022
Grant date	Feb 8, 2022

Abstract

A method of image processing is provided, the method comprising the steps of locating at least one object in an image of a scene, selecting at least a portion of the image of the scene in accordance with the location of the at least one object in the image of the scene, generating a different image of the at least one object in accordance with the selected portion of the image of the scene, the different image comprising the at least one object and being smaller than the image of the scene and detecting a plurality of points corresponding to parts of the at least one object located in the scene using the different image of the at least one object.

First claim

The invention claimed is:

1. A method of image processing, the method comprising:
locating at least one object in an image of a scene;
selecting at least a portion of the image of the scene in accordance with the location of the at least one object in the image of the scene;
generating a different image of the at least one object in accordance with the selected portion of the image of the scene, the different image comprising the at least one object and the different image being smaller than the image of the scene;
detecting a plurality of points corresponding to parts of the at least one object located in the scene using the different image of the at least one object;
wherein the image of the scene comprises a plurality of objects, and wherein generating the different image of the plurality of objects comprises generating a single different image of the plurality of objects, the different image comprising the plurality of objects in a predetermined configuration,
wherein selecting the at least one portion of the image of the scene in accordance with the location of the at least one object comprises selecting a region, or selecting an outside portion of the image outside the region, of the image of the scene of a predetermined size around the at least one object, and
wherein generating the different image of the scene comprises deleting the portion of the image of the scene outside the selected region of the image of the scene or deleting the selected outside portion of the image of the scene.

2. The method according to claim 1 , wherein locating the at least one object in the image of the scene comprises performing a foreground object extraction on the image of the scene.

3. The method according to claim 2 , wherein performing the foreground extraction comprises performing at least one of a chromatic distribution analysis of the image, a contrast distribution analysis of the image or a frame differencing with respect to a previous image of the scene.

4. The method according to claim 1 , wherein locating the at least one object in the image comprises comparing the image of the scene with a predetermined target image of the object.

5. The method according to claim 1 , wherein the different image of the at least one object is a cut-out of the at least one object from the image of the scene.

6. The method according to claim 1 , wherein prior to detecting a plurality of points, the method comprises magnifying the different image of the object to a predetermined size.

7. The method according to claim 1 , wherein prior to detecting a plurality of points, the method comprises performing a resolution adjustment on the different image in order to upscale or downscale the different image to a predetermined resolution.

8. The method according to claim 1 , wherein when selecting the portion of the at least one object, the method comprises storing the location of the at least one object in the original image in a storage unit.

9. The method according to claim 8 , comprising using the plurality of points corresponding to parts of the at least one object and the location of the at least one object in the image of the scene to replace the at least one object in the image of the scene with the corresponding plurality of points.

10. The method according to claim 8 , comprising storing an association between the plurality of points and the location of the at least one object in the storage unit.

11. The method according to claim 8 , comprising generating a resultant image of the plurality of points for each of the at least one object located in the scene, and storing an association between the resultant image of the plurality of points and the location of the object in the scene.

12. The method according to claim 1 , wherein the at least one object is a person and the plurality of points comprise points corresponding to at least one limb of the person.

13. The method according to claim 12 , wherein the method comprises generating a model for the at least one object in accordance with the plurality of points.

14. The method according to claim 13 , comprising generating a confidence score for the plurality of points in accordance with a biomechanical analysis of the model of the at least one object.

15. The method according to claim 14 , comprising generating a confidence score for the plurality of points in accordance with a second plurality of points obtained for that object from a second image of the scene; the second image of the scene having a different view of the at least one object in the scene.

16. The method according to claim 15 , wherein the image of the scene and the second image of the scene having a different view of the at least one object in the scene are obtained from different image capture devices located around the scene.

17. An apparatus for image processing, the apparatus comprising circuitry configured to:
locate at least one object in an image of a scene;
select at least a portion of the image of the scene in accordance with the location of the at least one object in the image of the scene;
generate a different image of the at least one object in accordance with the selected portion of the image of the scene, the different image comprising the at least one object and the different image being smaller than the image of the scene;
detect a plurality of points corresponding to parts of the at least one object located in the scene using the different image of the at least one object,
wherein the circuitry is further configured to select the at least one portion of the image of the scene in accordance with the location of the at least one object by being configured to select a region, or select an outside portion of the image outside the region, of the image of the scene of a predetermined size around the at least one object, and
wherein the circuitry is further configured to generate the different image of the scene by being configured to delete the portion of the image of the scene outside the selected region of the image of the scene or delete the selected outside portion of the image of the scene.

18. A non-transitory computer program product comprising instructions which, when the program is executed by a computer, cause the computer to cavy out a method of image processing, the method comprising:
locating at least one object in an image of a scene;
selecting at least a portion of the image of the scene in accordance with the location of the at least one object in the image of the scene;
generating a different image of the at least one object in accordance with the selected portion of the image of the scene, the different image comprising the at least one object and the different image being smaller than the image of the scene;
detecting a plurality of points corresponding to parts of the at least one object located in the scene using the different image of the at least one object,
wherein selecting the at least one portion of the image of the scene in accordance with the location of the at least one object comprises selecting a region, or selecting an outside portion of the image outside the region, of the image of the scene of a predetermined size around the at least one object, and
wherein generating the different image of the scene comprises deleting the portion of the image of the scene outside the selected region of the image of the scene or deleting the selected outside portion of the image of the scene.

Assignees

• Sony Corp
• Sony Europe Bv

Inventors

• Dodd Luke

Classifications

• G06T7/11 Primary

  Region-based segmentation · CPC title

• G06T2207/20132

  Image cropping · CPC title

• G06T7/251

  involving models · CPC title

• G06T2207/30196

  Human being; Person · CPC title

• G06T3/40

  Scaling of whole images or parts thereof, e.g. expanding or contracting · CPC title