Conversion of monoscopic visual content to stereoscopic 3D

1. A method comprising:

receiving a two-dimensional image to be converted to a first three-dimensional image;

identifying a second three-dimensional image sharing a characteristic with the two-dimensional image;

identifying a feature-to-depth mapping function associated with the second three-dimensional image in a data store, the data store storing feature-to-depth mapping functions for a plurality of three-dimensional images comprising the second three-dimensional image sharing the characteristic with the two-dimensional image, each stored feature-to-depth mapping function computed for a stereo image pair associated with a corresponding three-dimensional image of the plurality of three-dimensional images, and wherein the feature-to-depth mapping function associated with the second three-dimensional image associates a plurality of pixels in the second three-dimensional image having a first color value with a depth value;

determining a depth value for a plurality of pixels of the two-dimensional image according to the feature-to-depth mapping function; and

generating, by a processing device, the first three-dimensional image based on the depth value for the plurality of pixels of the two-dimensional image.

2. The method of claim 1 , further comprising:

identifying the second three-dimensional image using a K nearest neighbor algorithm.

3. The method of claim 1 , wherein the shared characteristic comprises a visual similarity.

4. The method of claim 1 , wherein each of the stored feature-to-depth mapping functions is associated with a different three-dimensional image.

5. The method of claim 4 , further comprising:

combining the plurality of stored feature-to-depth mapping functions to generate a final feature-to-depth mapping function.

6. The method of claim 1 , further comprising:

generating a depth map based on the determined depth value for the plurality of pixels of the two-dimensional image.

7. The method of claim 1 , further comprising:

identifying a plurality of feature-to-depth mapping functions associated with the second three-dimensional image, wherein each of the plurality of feature-to-depth mapping functions is associated with a different region of the second three-dimensional image.

8. A non-transitory machine-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising:

receiving a two-dimensional image to be converted to a first three-dimensional image;

identifying a second three-dimensional image sharing a characteristic with the two-dimensional image;

identifying a feature-to-depth mapping function associated with the second three-dimensional image in a data store, the data store storing feature-to-depth mapping functions for a plurality of three-dimensional images comprising the second three-dimensional image sharing the characteristic with the two-dimensional image, each stored feature-to-depth mapping function computed for a stereo image pair associated with a corresponding three-dimensional image of the plurality of three-dimensional images, and wherein the feature-to-depth mapping function associated with the second three-dimensional image associates a plurality of pixels in the second three-dimensional image having a first color value with a depth value;

determining a depth value for a plurality of pixels of the two-dimensional image according to the feature-to-depth mapping function; and

generating, by a processing device, the first three-dimensional image based on the depth value for the plurality of pixels of the two-dimensional image.

9. The non-transitory machine-readable storage medium of claim 8 , the operations further comprising:

identifying the second three-dimensional image using a K nearest neighbor algorithm.

10. The non-transitory machine-readable storage medium of claim 8 , wherein the shared characteristic comprises a visual similarity.

11. The non-transitory machine-readable storage medium of claim 8 , wherein each of the stored feature-to-depth mapping functions is associated with a different three-dimensional image.

12. The non-transitory machine-readable storage medium of claim 11 , the operations further comprising:

combining the plurality of stored feature-to-depth mapping functions to generate a final feature-to-depth mapping function.

13. The non-transitory machine-readable storage medium of claim 8 , the operations further comprising:

generating a depth map based on the determined depth value for the plurality of pixels of the two-dimensional image.

14. The non-transitory machine-readable storage medium of claim 8 , the operations further comprising:

identifying a plurality of feature-to-depth mapping functions associated with the second three-dimensional image, wherein each of the plurality of feature-to-depth mapping functions is associated with a different region of the second three-dimensional image.

15. A method comprising:

identifying a stereo image pair from a data store, the stereo image pair being associated with a first three-dimensional image;

calculating a disparity value for the stereo image pair;

determining, by a processing device, a plurality of depth values for the stereo image pair based on the disparity value;

generating, based on the plurality of depth values, a feature-to-depth mapping function for the stereo image pair, wherein the feature-to-depth mapping function associates a plurality of pixels in the stereo image pair having a first color value with one of the plurality of depth values; and

storing the feature-to-depth mapping function generated for the stereo image pair in association with the three-dimensional image in the data store storing feature-to-depth mapping functions for a plurality of three-dimensional images comprising the first three-dimensional image.

16. The method of claim 15 , wherein the disparity value comprises a horizontal shift of a point between a first image and a second image of the stereo image pair.

17. The method of claim 15 , further comprising:

calculating a hash for a first image of the stereo image pair; and

storing the hash and the feature-to-depth mapping function in association with each other in the data store.