Automatic generation of video from spherical content using audio/visual analysis

• US 9,652,667 B2
• Filed: 03/03/2015
• Issued: 05/16/2017
• Est. Priority Date: 03/04/2014
• Status: Active Grant

• Current Assignee: GoPro, Inc.
• Original Assignee: GoPro, Inc.
• Inventors: MacMillan, Timothy, Newman, David A., Adsumilli, Balineedu Chowdary
• Primary Examiner(s): Shibru, Helen
• Application Number:
  US14/637,193
• Publication Number:
  US 20150256746A1
• Time in Patent Office: 805 Days
• Field of Search: 386285, 382118, 348 36- 37, 348169
• CPC Class Codes: G03B 37/04 with cameras or projectors providing touching or overlapping fields of view
• CPC Class Codes: G06F 16/71 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06T 3/12 Panospheric to cylindrical image transformations
• CPC Class Codes: G06V 20/40 in video content extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44
• CPC Class Codes: H04L 65/612 for unicast
• CPC Class Codes: H04L 65/762 at the source reformatting of additional data in video distribution servers H04N21/2355
• CPC Class Codes: H04N 13/106 Processing image signals for multi-view video sequence encoding H04N19/597
• CPC Class Codes: H04N 21/233 Processing of audio elementary streams monitoring, identification or recognition of audio in broadcast systems H04H60/58
• CPC Class Codes: H04N 21/23418 involving operations for analysing video streams, e.g. detecting features or characteristics television picture signal circuitry for scene change detection H04N5/147; filtering for image enhancement G06T5/00; methods or arrangements for recognising scenes G06V20/00; arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59
• CPC Class Codes: H04N 23/698 for achieving an enlarged field of view, e.g. panoramic image capture

Abstract

A spherical content capture system captures spherical video content. A spherical video sharing platform enables users to share the captured spherical content and enables users to access spherical content shared by other users. In one embodiment, captured metadata or video/audio processing is used to identify content relevant to a particular user based on time and location information. The platform can then generate an output video from one or more shared spherical content files relevant to the user. The output video may include a non-spherical reduced field of view such as those commonly associated with conventional camera systems. Particularly, relevant sub-frames having a reduced field of view may be extracted from each frame of spherical video to generate an output video that tracks a particular individual or object of interest.

Claims

1. A method for generating an output video from spherical video content, the method comprising:

storing, by a video server, a first spherical video obtained from a first camera system comprising a first sequence of spherical video frames, each having a first spherical field of view

storing, by a video server, a second spherical video obtained from a second camera system comprising a second sequence of spherical video frames, each having a second spherical field of view;

processing, by the video server, the first spherical video to identify a target audio or visual feature of interest meeting one or more audio or visual criteria;

determining, by the video server, a first range of frames of the first spherical video having the target feature of interest;

determining, by the video server, a second range of frames of the second spherical video having the target feature of interest;

determining, by the video server, a first sequence of sub-frames from each of the first range of frames, each of the first sequence of sub-frames having a non-spherical field of view, and each of the first sequence of sub-frames including a spatial region around the target feature of interest;

determining, by the video server, a second sequence of sub-frames from each of the second range of frames, each of the second sequence of sub-frames having a non-spherical field of view, and each of the second sequence of sub-frames including a spatial region around the target feature of interest;

generating, by the video server, a first combined sequence of sub-frames including the target feature of interest, the combined sequence of sub-frames comprising the first sequence of sub-frames and the second sequence of sub-frames;

generating, by the video server, a first portion of an output video including the first combined sequence of sub-frames; and

outputting the output video.

2. The method of claim 1, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a facial recognition algorithm on the first spherical video to identify first spatial locations of one or more faces depicted in the first spherical video;

performing the facial recognition algorithm on the second spherical video to identify second spatial locations of the one or more faces depicted in the second spherical video.

3. The method of claim 1, wherein processing the first spherical video to identify the target feature comprises:

performing an object recognition algorithm on the first spherical video to identify first spatial locations of one or more objects depicted in the first spherical; and

performing the object recognition algorithm on the second spherical video to identify second spatial locations of one or more objects depicted in the first spherical video.

4. The method of claim 1, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a motion analysis algorithm on the first spherical video to identify first spatial locations of an object depicted in the first spherical video meeting predefined motion parameters;

performing the motion analysis algorithm on the second spherical video to identify second spatial locations of the object depicted in the second spherical video meeting the predefined motion parameters.

5. The method of claim 1, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a gesture recognition algorithm on the first spherical video to identify first spatial locations of an individual performing a particular gesture depicted in the first spherical video; and

performing the gesture recognition algorithm on the second spherical video to identify second spatial locations of the individual performing the particular gesture depicted in the second spherical video.

6. The method of claim 1, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing an audio analysis on an audio track of first spherical video to identify a first direction between an audio source and the first camera capturing the first spherical audio;

performing the audio analysis on an audio track of the second spherical video to identify a second direction between the audio source and the second camera capturing the second spherical audio;

identifying a first spatial location of the audio source in the first spherical video based on the first direction between the audio source and the first camera; and

identifying a second spatial location of the audio source in the second spherical video based on the second direction between the audio source and the second camera.

7. The method of claim 1, further comprising:

storing, by the video server, a plurality of spherical videos comprising a sequence of spherical video frames each having a spherical field of view and one or more audio tracks obtained from a plurality of devices;

processing, by the video server, the plurality of spherical videos to identify the target feature of interest meeting the one or more audio or visual criteria;

determining, by the video server, a manifold range of frames of the plurality of spherical videos having the target feature of interest;

determining, by the video server, a manifold sequence of sub-frames from each of the manifold range of frames, each of the manifold sequence of sub-frames having a non-spherical field of view, and each of the manifold sequence of sub-frames including a spatial region around the target feature of interest;

determining, by the video server, a second combined sequence of sub-frames including the target feature of interest comprised of the first combination of sub-frames and the manifold sequence of sub-frames;

generating, by the video server, a second portion of the output video including the second combination of sub-frames; and

outputting the output video.

8. A non-transitory computer-readable storage medium storing instructions for generating an output video from spherical video content, the instructions when executed by one or more processors causing the one or more processors to perform steps including:

storing a first spherical video obtained from a first camera system comprising a first sequence of spherical video frames each having a first spherical field of view;

storing a second spherical video obtained from a second camera system comprising a second sequence of spherical video frames, each having a second spherical field of view;

processing the first spherical video to identify a target audio or visual feature of interest meeting one or more audio or visual criteria;

determining a first range of frames of the first spherical video having the target feature of interest;

determining a second range of frames of the second spherical video having the target feature of interest;

determining a first sequence of sub-frames from each of the first range of frames, each of the first sequence of sub-frames having a non-spherical field of view, and each of the first sequence of sub-frames including a spatial region around the target feature of interest;

determining a second sequence of sub-frames from each of the second range of frames, each of the second sequence of sub-frames having a non-spherical field of view, and each of the second sequence of sub-frames including a spatial region around the target feature of interest;

generating a first combined sequence of sub-frames including the target feature of interest comprising of the first sequence of sub-frames and the second sequence of sub-frames;

generating a first portion of an output video including the first combined sequence of sub-frames; and

outputting the output video.

9. The non-transitory computer-readable storage medium of claim 8, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a facial recognition algorithm on the first spherical video to identify first spatial locations of one or more faces depicted in the first spherical video;

performing the facial recognition algorithm on the second spherical video to identify second spatial locations of the one or more faces depicted in the second spherical video.

10. The non-transitory computer-readable storage medium of claim 8, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing an object recognition algorithm on the first spherical video to identify spatial locations of one or more objects depicted in the first spherical video; and

performing the object recognition algorithm on the second spherical video to identify second spatial locations of the one or more objects depicted in the second spherical video.

11. The non-transitory computer-readable storage medium of claim 8, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a motion analysis algorithm on the first spherical video to identify spatial locations of an object depicted in the first spherical video meeting predefined motion parameters; and

performing the motion analysis algorithm on the second spherical video to identify second spatial locations of the object depicted in the second spherical video meeting predefined motion parameters.

12. The non-transitory computer-readable storage medium of claim 8, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a gesture recognition algorithm on the first spherical video to identify spatial locations of an individual performing a particular gesture depicted in the first spherical video;

performing the gesture recognition algorithm on the second spherical video to identify second spatial locations of the individual performing a particular gesture depicted in the second spherical video.

13. The non-transitory computer-readable storage medium of claim 8, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing an audio analysis on an audio track of first spherical video to identify a first direction between an audio source and the first camera capturing the first spherical audio;

performing the audio analysis on an audio track of the second spherical video to identify a second direction between the audio source and the second camera capturing the second spherical audio;

identifying a first spatial location of the audio source in the first spherical video based on the first direction between the audio source and the first camera; and

identifying a second spatial location of the audio source in the second spherical video based on the second direction between the audio source and the second camera.

14. The non-transitory computer-readable storage medium of claim 8, wherein the instructions when executed by the one or more processors further cause the one or more processors to perform steps including:

storing a plurality of spherical videos comprising a sequence of spherical video frames each having a spherical field of view and one or more audio tracks obtained from a plurality of devices;

processing the plurality of spherical videos to identify the target feature of interest meeting the one or more audio or visual criteria;

determining a manifold range of frames of the plurality of spherical videos having the target feature of interest;

determining a manifold sequence of sub-frames from each of the manifold range of frames, each of the manifold sequence of sub-frames having a non-spherical field of view, and each of the manifold sequence of sub-frames including a spatial region around the target feature of interest; and

determining a second combination of sub-frames including the target feature of interest comprised of the first combination of sub-frames and the manifold sequence of sub-frames; and

generating a second portion of the output video including the second sequence of sub-frames; and

outputting the output video.

15. A video server for generating an output video from spherical video content, the video server comprising:

one or more processors; and

a non-transitory computer-readable storage medium storing instructions that when executed by the one or more processors causing the one or more processors to perform steps including;

storing a first spherical video from a first camera system comprising a first sequence of spherical video frames each having a first spherical field of view;

storing a second spherical video from a second camera system comprising a second sequence of spherical video frames, each having a second spherical field of view;

processing the first spherical video to identify a target audio or visual feature of interest meeting one or more audio or visual criteria;

determining a first range of frames of the first spherical video having the target feature of interest;

determining, by the video server, a second range of frames of the second spherical video having the target feature of interest;

determining a first sequence of sub-frames from each of the first range of frames, each of the first sequence of sub-frames having a non-spherical field of view, and each of the first sequence of sub-frames including a spatial region around the target feature of interest;

determining a second sequence of sub-frames from each of the second range of frames, each of the second sequence of sub-frames having a non-spherical field of view, and each of the second sequence of sub-frames including a spatial region around the target feature of interest;

generating a first combined sequence of sub-frames including the target feature of interest comprising of the first sequence of sub-frames and the second sequence of sub-frames;

generating a first portion of an output video including the first combined sequence of sub-frames; and

outputting the output video.

16. The video server of claim 15, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing a facial recognition algorithm on the first spherical video to identify first spatial locations of one or more faces depicted in the first spherical video

performing the facial recognition algorithm on the second spherical video to identify second spatial locations of the one or more faces depicted in the first spherical video.

17. The video server of claim 15, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing an object recognition algorithm on the first spherical video to identify first spatial locations of one or more objects depicted in the first spherical video;

performing the object recognition algorithm on the second spherical video to identify second spatial locations of the one or more objects depicted in the first spherical video.

18. The video server of claim 15, wherein determining the first and the second sequence of sub-frames including the target feature comprises:

performing an audio analysis on an audio track of first spherical video to identify a first direction between an audio source and the first camera capturing the first spherical audio;

performing the audio analysis on an audio track of the second spherical video to identify a second direction between the audio source and the second camera capturing the second spherical audio;

identifying a first spatial location of the audio source in the first spherical video based on the first direction between the audio source and the first camera; and

identifying a second spatial location of the audio source in the second spherical video based on the second direction between the audio source and the second camera.

19. The video server of claim 15, determining the first and the second sequence of sub-frames including the target feature comprises:

performing a motion analysis algorithm on the first spherical video to identify spatial locations of an object depicted in the first spherical video meeting predefined motion parameters; and

performing the motion analysis algorithm on the second spherical video to identify second spatial locations of the object depicted in the second spherical video meeting predefined motion parameters.