Cinematic space-time view synthesis for enhanced viewing experiences in computing environments

1. An apparatus comprising:

one or more processors coupled to one or more cameras, the one or more processors to:

facilitate the one or more cameras to capture multiple images at multiple positions or multiple points in times, wherein the multiple images represent multiple views of an object or a scene;

generate a displacement map of the multiple images based on movement of pixels from a first image of the multiple images to a second image of the multiple images, wherein the displacement map represents a flow to a target view of the multiple views such that to track one or more pixels from the first image moving into the second image, wherein the first and second images are warped up to half-way to generate a middle image, wherein warping of the first and second images include warping of pixels of the first and second images based on the displacement map; and

synthesize the multiple images into a single image based on the displacement map, wherein the single image includes a final middle image representing an intermediary view of the multiple views, wherein the final middle image represents a final output view of the multiple views.

2. The apparatus of claim 1 , wherein the multiple images comprise at least one of two space images or two time images, wherein the two space images are captured by at least two cameras at a single point in time, wherein the two time images are captured by a single camera at two points in time.

3. The apparatus of claim 1 , wherein the middle image comprises at least one of a middle space image of the two space images, a middle time image of the two time images, and a middle space-time image of the multiple images.

4. The apparatus of claim 1 , wherein the one or more processors are further to facilitate a warping layer to warp the multiple images into warped images based on the displacement map, wherein the multiple images are synthesized into the single image based on the warped images.

5. The apparatus of claim 1 , wherein the one or more processors are further to train a neural network, wherein training includes end-to-end training facilitating access to additional training data if the neural network serving as a main network is segmented into sub-components, wherein the neural network comprises a convolutional neutral network (CNN).

6. The apparatus of claim 1 , wherein the one or more processors comprise a graphics processor co-located with an application processor on a common semiconductor package.

7. A method comprising:

capturing, by one or more cameras coupled to one or more processors, multiple images at multiple positions or multiple points in times, wherein the multiple images represent multiple views of an object or a scene;

generating a displacement map of the multiple images based on movement of pixels from a first image of the multiple images to a second image of the multiple images, wherein the displacement map represents a flow to a target view of the multiple views such that to track one or more pixels from the first image moving into the second image, wherein the first and second images are warped up to half-way to generate a middle image, wherein warping of the first and second images include warping of pixels of the first and second images based on the displacement map; and

synthesizing, by a neural network, the multiple images into a single image based on the displacement map, wherein the single image includes a final middle image representing an intermediary view of the multiple views, wherein the final middle image represents a final output view of the multiple views.

8. The method of claim 7 , wherein the multiple images comprise at least one of two space images or two time images, wherein the two space images are captured by at least two cameras at a single point in time, wherein the two time images are captured by a single camera at two points in time.

9. The method of claim 7 , wherein the middle image comprises at least one of a middle space image of the two space images, a middle time image of the two time images, and a middle space-time image of the multiple images.

10. The method of claim 7 , further comprising facilitating, by a custom layer of the neural network, a warping layer to warp the multiple images into warped images based on the displacement map, wherein the multiple images are synthesized into the single image based on the warped images.

11. The method of claim 7 , further comprising training the neural network, wherein training includes end-to-end training facilitating access to additional training data if the neural network serving as a main network is segmented into sub-components, wherein the neural network comprises a convolutional neutral network (CNN).

12. The method of claim 7 , wherein the one or more processors comprise a graphics processor co-located with an application processor on a common semiconductor package.

13. At least one non-transitory machine-readable medium comprising instructions which, when executed by a computing device, cause the computing device to perform operations comprising:

capturing, by one or more cameras coupled to one or more processors, multiple images at multiple positions or multiple points in times, wherein the multiple images represent multiple views of an object or a scene; and

generating a displacement map of the multiple images based on movement of pixels from a first image of the multiple images to a second image of the multiple images, wherein the displacement map represents a flow to a target view of the multiple views such that to track one or more pixels from the first image moving into the second image, wherein the first and second images are warped up to half-way to generate a middle image, wherein warping of the first and second images include warping of pixels of the first and second images based on the displacement map; and

synthesizing, by a neural network, the multiple images into a single image based on the displacement map, wherein the single image includes a final middle image representing an intermediary view of the multiple views, wherein the final middle image represents a final output view of the multiple views.

14. The non-transitory machine-readable medium of claim 13 , wherein the multiple images comprise at least one of two space images or two time images, wherein the two space images are captured by at least two cameras at a single point in time, wherein the two time images are captured by a single camera at two points in time.

15. The non-transitory machine-readable medium of claim 13 , wherein the middle image comprises at least one of a middle space image of the two space images, a middle time image of the two time images, and a middle space-time image of the multiple images.

16. The non-transitory machine-readable medium of claim 13 , wherein the operations further comprise facilitating, by a custom layer of the neural network, a warping layer to warp the multiple images into warped images based on the displacement map, wherein the multiple images are synthesized into the single image based on the warped images.

17. The non-transitory machine-readable medium of claim 13 , wherein the operations further comprise training the neural network, wherein training includes end-to-end training facilitating access to additional training data if the neural network serving as a main network is segmented into sub-components, wherein the neural network comprises a convolutional neutral network (CNN), wherein the one or more processors comprise a graphics processor co-located with an application processor on a common semiconductor package.