Methods and systems of creating a stereoscopic image
View Patent ↗Disclosed herein are methods and systems for creating stereoscopic images. A left-eye view image for a stereoscopic image and an imperfect right-eye view image may be received. A smooth optical flow may be generated from the received image representing the left-eye view to the imperfect image representing the right-eye view to produce a first candidate right-eye view. Objects are identified in the imperfect image, and merged into the first candidate right-eye view image to create a right-eye view image. A stereoscopic image is created from the left-eye view image and right-eye view image.
1. A computer-implemented method, comprising:
generating, with at least one processing device, a smooth optical flow from a first image to a second image, wherein the first image represents one of a left-eye view and a right-eye view, and the second image represents the other one of a left-eye view and a right-eye view, and wherein the second image is an imperfect image that has at least one of a feature point having distorted height, a discontinuity in depth, an object with distorted shape, or an object with distorted appearance;
producing a first candidate view from the smooth optical flow;
identifying objects in the second image;
merging the identified objects of the second image into the first candidate view;
creating a third image based on the merging; and
creating a target stereoscopic image from the first image and the third image.
2. The method of claim 1 , wherein the first image represents a left-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a right-eye view of the target stereoscopic image.
3. The method of claim 1 , wherein the first image represents a right-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a left-eye view of the target stereoscopic image.
4. The method of claim 1 , wherein the step of identifying objects in the second image is performed by a clustering algorithm.
5. The method of claim 4 , wherein the clustering algorithm is K-Means.
6. The method of claim 1 , wherein the step of merging the identified objects of the second image into the first image reduces artifacts and increases detail.
7. The method of claim 1 , wherein the smooth optical flow is generated by interpolating vectors between matching pixels in the first image and the second image.
8. The method of claim 1 , further comprising producing a first candidate view from the smooth optical flow using computer vision techniques.
9. The method of claim 1 , further comprising merging the identified objects of the second image into the first candidate view using Poisson image editing.
10. A non-transitory computer readable storage medium containing control logic stored thereon that, when executed by one or more processing devices, causes the one or more processing devices to:
generate a smooth optical flow from a first image to a second image, wherein the first image represents one of a left-eye view and a right-eye view, and the second image represents the other one of a left-eye view and a right-eye view, and wherein the second image is an imperfect image that has at least one of a feature point having distorted height, a discontinuity in depth, an object with distorted shape, or an object with distorted appearance;
produce a first candidate view from the smooth optical flow;
identify objects in the imperfect image;
merge the identified objects of the second image into the first candidate view;
create a third image based on the merging; and
create a target stereoscopic image from the first image and the second image.
11. The computer readable storage medium of claim 10 , wherein the first image represents a left-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a right-eye view of the target stereoscopic image.
12. The computer readable storage medium of claim 10 , wherein the first image represents a right-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a left-eye view of the target stereoscopic image.
13. The computer readable storage medium of claim 10 , wherein the control logic causes the one or more processing devices to identify objects in the second image by a clustering algorithm.
14. The computer readable storage medium of claim 13 wherein the clustering algorithm is K-Means.
15. The computer readable storage medium of claim 10 , wherein the smooth optical flow is generated by interpolating vectors between matching pixels in the first image and the second image.
16. The computer readable storage medium of claim 10 , wherein the control logic causes the one or more processing devices to produce a first candidate view from the smooth optical flow by computer vision techniques.
17. The computer readable storage medium of claim 10 , wherein the control logic causes the one or more processing devices to merge the identified objects of the second image into the first candidate view using Poisson image editing.
18. A system, comprising:
a receiver that receives a first and second image, wherein the first image represents one of a left-eye view and a right-eye view, and the second image represents the other one of a left-eye view and a right-eve view, and wherein the second image is an imperfect image that has at least one of a feature point having distorted height, a discontinuity in depth, an object with distorted shape, or an object with distorted appearance;
a smooth optical flow generator that generates a smooth optical flow from the first image to the second image;
a candidate image producer that produces a first candidate view from the smooth optical flow;
an object identifier that identifies objects in the second image;
an image creator that creates a third image by merging objects identified by the object identifier into the first candidate view; and
a stereoscopic image creator, implemented on at least on processing device, that creates a stereoscopic image from the first image and the third image.
19. The system of claim 18 , wherein the first image represents a left-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a right-eye view of the target stereoscopic image.
20. The system of claim 18 , wherein the first image represents a right-eye view of a target stereoscopic image, and wherein the second image is an imperfect image representing a left-eye view of the target stereoscopic image.
21. The system of claim 18 , wherein the object identifier further identifies objects in the second image using K-Means clustering.
22. The system of claim 18 , wherein the image creator merges objects identified by the object identifier into the first candidate view using Poisson image editing.