Predictive light field compression

1. A method for manipulating a light-field image, the method comprising:

in a data store, receiving the light-field image after the light-field image has been captured by a light-field image capture device;

in a processor, generating a first refocus image pool comprising a plurality of refocus images based on the light-field image; and

in the processor, using the first refocus image pool to compress the light-field image to generate a bitstream, smaller than the light-field image, that is representative of the light-field image;

wherein using the first refocus image pool to compress the light-field image comprises:

dividing the light-field image into a plurality of blocks;

for each of the blocks, using the first refocus image pool to generate a compressed block by:

generating a prediction for each of the refocus images; and

comparing the generated predictions to select a best prediction believed to be most representative of the block; and

combining the compressed blocks to generate the bitstream.

2. The method of claim 1 , wherein generating the first refocus image pool comprises projecting the light-field image at one or more planes.

3. The method of claim 1 , further comprising, based on results of comparing the generated predictions, updating the plurality of refocus images.

4. The method of claim 1 , wherein comparing the generated predictions is performed using an algorithm selected from the group consisting of least-distortion-error selection, least-reconstruction-error selection, and minimal-rate-distortion selection.

5. The method of claim 1 , wherein the light-field image comprises a first light-field image from a first time index in light-field video, the method further comprising:

in the data store, receiving a second light-field image from a second time index in the light-field video; and

in the processor, generating a second refocus image pool comprising a second plurality of refocus images based on the second light-field image;

wherein using the first refocus image pool to compress the light-field image to generate the bitstream further comprises using the second refocus image pool to facilitate generation of the bitstream.

6. The method of claim 1 , further comprising:

in one of the data store and a second data store, receiving the bitstream;

in one of the processor and a second processor, generating a second refocus image pool comprising a second plurality of refocus images based on the bitstream; and

in one of the processor and the second processor, using the second refocus image pool to decompress the bitstream to generate a reconstructed light-field image, larger than the bitstream, that is representative of the light-field image.

7. The method of claim 1 , wherein the light-field image capture device comprises a plenoptic light-field camera comprising image sensor and a microlens array.

8. A method for manipulating a light-field image, the method comprising:

in a first data store, receiving the light-field image after the light-field image has been captured by a light-field image capture device;

in a first processor, generating a first refocus image pool comprising a plurality of refocus images based on the light-field image;

in the processor, using the first refocus image pool to compress the light-field image to generate a bitstream, smaller than the light-field image, that is representative of the light-field image;

in one of the first data store and a second data store, receiving the bitstream;

in one of the first processor and a second processor, generating a second refocus image pool comprising a second plurality of refocus images based on the bitstream; and

in one of the first processor and the second processor, using the second refocus image pool to decompress the bitstream to generate a reconstructed light-field image, larger than the bitstream, that is representative of the light-field image;

wherein using the second refocus image pool to decompress the bitstream comprises:

dividing the bitstream into a plurality of blocks;

for each of the blocks, using the second refocus image pool to generate a reconstructed block by:

generating a prediction based on at least one of the second plurality of refocus images; and

applying at least one selection from the group consisting of inverse transformation, inverse quantization, and entropy decoding; and

combining the reconstructed blocks to generate the reconstructed light-field image.

9. The method of claim 8 , further comprising, based on the prediction, updating the second plurality of refocus images.

10. The method of claim 8 , wherein the light-field image capture device comprises a plenoptic light-field camera comprising image sensor and a microlens array.

11. A method for generating a reconstructed light-field image, the method comprising:

in a data store, receiving a bitstream generated by compressing a light-field image;

in a processor, generating a refocus image pool comprising a plurality of refocus images based on the bitstream; and

in the processor, using the refocus image pool to decompress the bitstream to generate the reconstructed light-field image such that the reconstructed light-field image is larger than the bitstream and is representative of the light-field image;

wherein using the refocus image pool to decompress the bitstream comprises:

dividing the bitstream into a plurality of blocks;

for each of the blocks, using the refocus image pool to generate a reconstructed block by generating a prediction based on at least one of the plurality of refocus images; and

combining the reconstructed blocks to generate the reconstructed light-field image.

12. The method of claim 11 , further comprising, based on the prediction, updating the plurality of refocus images.

13. A non-transitory computer-readable medium for manipulating a light-field image, comprising instructions stored thereon, that when executed by a processor, perform the steps of:

causing a data store to receive the light-field image after the light-field image has been captured by a light-field image capture device;

generating a first refocus image pool comprising a plurality of refocus images based on the light-field image; and

using the first refocus image pool to compress the light-field image to generate a bitstream, smaller than the light-field image, that is representative of the light-field image;

wherein using the first refocus image pool to compress the light-field image comprises:

dividing the light-field image into a plurality of blocks;

for each of the blocks, using the first refocus image pool to generate a compressed block by:

generating a prediction for each of the refocus images; and

comparing the generated predictions to select a best prediction believed to be most representative of the block; and

combining the compressed blocks to generate the bitstream.

14. The non-transitory computer-readable medium of claim 13 , wherein generating the first refocus image pool comprises projecting the light-field image at one or more planes.

15. The non-transitory computer-readable medium of claim 13 , further comprising instructions stored thereon, that when executed by a processor, perform the step of, based on results of comparing the generated predictions, updating the plurality of refocus images.

16. The non-transitory computer-readable medium of claim 13 , further comprising instructions stored thereon, that when executed by a processor, perform the steps of:

causing one of the data store and a second data store to receive the bitstream;

generating a second refocus image pool comprising a second plurality of refocus images based on the bitstream; and

using the second refocus image pool to decompress the bitstream to generate a reconstructed light-field image, larger than the bitstream, that is representative of the light-field image.

17. A non-transitory computer-readable medium for generating a reconstructed light-field image, comprising instructions stored thereon, that when executed by a processor, perform the steps of:

causing a data store to receive a bitstream generated by compressing a light-field image;

generating a refocus image pool comprising a plurality of refocus images based on the bitstream; and

using the refocus image pool to decompress the bitstream to generate the reconstructed light-field image such that the reconstructed light-field image is larger than the bitstream and is representative of the light-field image;

wherein using the refocus image pool to decompress the bitstream comprises:

dividing the bitstream into a plurality of blocks;

for each of the blocks, using the refocus image pool to generate a reconstructed block by generating a prediction based on at least one of the plurality of refocus images; and

combining the reconstructed blocks to generate the reconstructed light-field image.

18. The non-transitory computer-readable medium of claim 17 , further comprising instructions stored thereon, that when executed by a processor, perform the step of, based on the prediction, updating the plurality of refocus images.

19. A system for manipulating a light-field image, the system comprising:

a data store configured to receive the light-field image after the light-field image has been captured by a light-field image capture device; and

a processor, communicatively coupled to the data store, configured to:

generate a first refocus image pool comprising a plurality of refocus images based on the light-field image; and

use the first refocus image pool to compress the light-field image to generate a bitstream, smaller than the light-field image, that is representative of the light-field image;

wherein the processor is configured to use the first refocus image pool to compress the light-field image by:

dividing the light-field image into a plurality of blocks;

for each of the blocks, using the first refocus image pool to generate the compressed block by:

generating a plurality of predictions by generating one of the predictions for each of the refocus images; and

comparing the predictions to select a best prediction believed to be most representative of the block for which the compressed block is generated; and

combining the compressed blocks to generate the bitstream.

20. The system of claim 19 , wherein the processor is configured to generate the first refocus image pool by projecting the light-field image at one or more planes.

21. The system of claim 19 , wherein the processor is further configured, based on results of comparing the predictions, to update the plurality of refocus images.

22. The system of claim 19 , wherein:

the data store or a second data store of the system is configured to receive the bitstream; and

the processor or a second processor of the system is configured to:

generate a second refocus image pool comprising a second plurality of refocus images based on the bitstream; and

use the second refocus image pool to decompress the bitstream to generate a reconstructed light-field image, larger than the bitstream, that is representative of the light-field image.

23. A system for generating a reconstructed light-field image, the system comprising:

a data store configured to receive a bitstream generated by compressing a light-field image; and

a processor, communicatively coupled to the data store, configured to:

generate a refocus image pool comprising a plurality of refocus images based on the bitstream; and

use the refocus image pool to decompress the bitstream to generate the reconstructed light-field image such that the reconstructed light-field image is larger than the bitstream and is representative of the light-field image;

wherein the processor is configured to use the refocus image pool to decompress the bitstream by:

dividing the bitstream into a plurality of blocks;

for each of the blocks, using the refocus image pool to generate a reconstructed block by generating a prediction based on at least one of the plurality of refocus images; and

combining the reconstructed blocks to generate the reconstructed light-field image.

24. The system of claim 23 , wherein the processor is further configured, based on the prediction, to update the plurality of refocus images.