Scene capture for reconstruction of obscured views
An imagery processing system determines alternative pixel color values for pixels of captured imagery where the alternative pixel color values are obtained from alternative sources. A main imagery capture device, such as a camera, captures main imagery such as still images and/or video sequences, of a live action scene. Alternative devices capture imagery of the live action scene, in some spectra and form, and that alternative imagery is processed to provide user-selectable alternatives for pixel ranges from the main imagery.
1. A computer-implemented method of processing images from a main imaging device using capture device inputs from capture devices, the method comprising:
obtaining a main image from a main image camera at a first position, wherein the main image comprises a main pixel array derived from a view of a live action scene;
obtaining a plurality of alternative images from one or more alternative cameras at positions that are uncoordinated with the main camera's first position, wherein each alternative image in the plurality of images comprises an alternative array of pixels including volume information;
determining positions and orientations for the alternative cameras; and
transferring the main image, alternative images, alternative camera positions and orientations for use in mapping pixels of the alternative array of pixels to pixels of the main pixel array.
2. The method of claim 1 , wherein mapping pixels of the alternative array of pixels to pixels of the main pixel array comprises, for at least one of the alternative array of pixels, translating, rotating, and/or shearing the pixel array of the at least one of the alternative array of pixels to align pixels with the main pixel array.
3. The method of claim 1 , wherein the set of one or more replacement regions comprises a single replacement region that forms a continuous region.
4. The method of claim 1 , wherein the set of one or more replacement regions comprises a plurality of discontinuous replacement regions.
5. The method of claim 1 , further comprising determining the set of one or more replacement regions using image segmentation to identify boundaries of objects appearing in the main pixel array.
6. The method of claim 1 , further comprising obtaining user input for determining the set of one or more replacement regions.
7. The method of claim 1 , wherein selecting one or more corresponding alternative regions includes selecting a plurality of corresponding alternative regions, the method further comprising:
blending the plurality of corresponding alternative regions to derive the corresponding alternative regions of the one or more selected alternative images for the replacement regions.
8. The method of claim 1 , wherein selecting one or more corresponding alternative regions comprises determining, using a voting or weighting process, which of a plurality of corresponding alternative regions provides a best fit for the replacement regions.
9. The method of claim 8 , further comprising determining the best fit for the replacement regions using a color matching process.
10. The method of claim 9 , wherein the color matching process comprises a color matching process across a plurality of frames of a main image device video sequence.
11. The method of claim 9 , wherein the color matching process comprises a color matching process across the main pixel array.
12. The method of claim 1 , further comprising:
providing one or more alternative regions to a user for selection.
13. The method of claim 1 , further comprising
providing one or more alternative regions to an automated process for selection.
14. The method of claim 1 , further comprising:
transferring the main image, alternative images, alternative camera positions and orientations to a reconstructor;
using the reconstructor to replace portions of a scene that were obstructed by objects so as to appear as if those removed objects were not present in the scene when it was captured.
15. The method of claim 14 , wherein the reconstructor uses machine-learning.
16. The method of claim 15 , wherein the reconstructor is trained on video with known values for what should be reconstructed.
17. An apparatus including a processor configured to perform the actions recited in claim 1 .
18. One or more non-transitory processor-readable media including instructions executable by one or more processors to perform the actions recited in claim 1 .