Video system for real-time panoramic video delivery

1. A method for processing video, comprising;

capturing video at a first location from multiple sensors;

stitching images corresponding to the multiple sensors to produce a panoramic video having at least a 100° field of view;

processing the panoramic video at cloud-based server for delivery to a plurality of client devices, wherein the processing comprises optimizing the panoramic video based on a feedback received from each of the plurality of client devices; and

streaming the processed video to the plurality of client devices.

2. The method of claim 1 , wherein the capturing and the stitching is by a multiple imager panoramic camera and is performed at a latency of less than 1.2 times a frame-time for rendered video.

3. The method of claim 1 , wherein the cloud-based server processes the video with a latency of less than 0.4 times a frame-time for rendered video.

4. The method of claim 1 , wherein the processing by the cloud-based server comprises customizing the panoramic video on a stream-by-stream basis, wherein each customized stream is customized for a particular client device based on region of interest (ROI) associated with said client device.

5. The method of claim 4 , further comprising determining the region of interest (ROI) for each client device based on at least one of zooming and panning information received from the client device.

6. The method of claim 4 , wherein each customized stream includes more bits in the video stream for areas corresponding to the region of interest (ROI) associated with that client device for that stream.

7. The method of claim 1 , wherein the stitching is based on a non-linear seam selected from a region of overlap between two adjacent images.

8. The method of claim 1 , wherein the stitching is based on a seam calculated between two adjacent images based on a technique that minimizes the temporal and spatial distortion seen when traversing the adjacent images in the region of overlap between the two images from top to bottom.

9. The method of claim 1 , further comprising:

encoding the panoramic video by determining a number of slices into which to slice each input frame of the panoramic video;

determining a number of stacks associated with an output encoding; and

slicing each input frame into the number of slices; and

stacking each slice into encoded frames of the output encoding.

10. A system for video capture and delivery, comprising:

a camera with multiple sensors to capture video at a first location, and to stitch images corresponding to the multiple sensors to produce a panoramic video having at least a 100° field of view; and

a cloud-based server to process the panoramic video for delivery to a plurality of client devices, and to stream the processed video to the plurality of client devices, wherein the processing comprises optimizing the panoramic video based on a feedback received from each of the plurality of client devices.

11. The system of claim 10 , wherein the capturing and the stitching is performed at a latency of less than 1.2 times a frame-time of rendered video.

12. The system of claim 10 , wherein the cloud-based server processes the video with a latency of less than 0.4 times a frame-time of rendered video.

13. The system of claim 10 , wherein the processing by the cloud-based server comprises customizing the panoramic video on a stream-by stream basis, wherein each customized stream is customized for a particular client device based on region of interest (ROI) associated with said client device.

14. The system of claim 10 , wherein the cloud-based server determines the region of interest (ROI) for each client device based on at least one of zooming and panning information received from the client device.

15. The system of claim 10 , wherein each customized stream includes more bits in the video stream for areas corresponding to the region of interest (ROI) associated with that client device for that stream.

16. The system of claim 10 , wherein the stitching is based on a non-linear seam selected from a region of overlap between two adjacent images.

17. The system of claim 10 , wherein the stitching is based on a seam calculated between two adjacent images based on a technique that minimizes the temporal and spatial distortion seen when traversing the adjacent images in region of overlap between the two images from top to bottom.

18. The system of claim 10 , wherein the camera encodes the panoramic video based on a technique comprising determining a number of slices into which to slice each input frame of the panoramic video; determining a number of stacks associated with an output encoding; slicing each input frame into the number of slices; and stacking each slice into encoded frames of the output encoding.