Media Summarization

Patent metadata

Field	Value
Publication number	US-9972356-B2
Application number	US-201615393359-A
Country	US
Kind code	B2
Filing date	Dec 29, 2016
Priority date	Sep 10, 2012
Publication date	May 15, 2018
Grant date	May 15, 2018

Abstract

Techniques for summarizing media are described. A viewer-interaction analyzer receives a media file containing media, the media file including a plurality of segments. A segment of the media file is scored based on interactions of a set of raters. Viewer metrics on the segment of the media file are measured based on interactions with the segment of the media file by a set of viewers. A set of feature vectors are formed based on the measured viewer interactions, where feature vectors in the set of feature vectors are based on interactions of the set of viewers. A model is trained based on the set of feature vectors and the score assigned to the segment of the media file. The model is applied to segments of the media file to generate an interest rating for segments of the media file. An edited media file is generated based on segments of the media file having interest ratings that meet a criterion. The Abstract is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims.

First claim

What is claimed is:

1. At least one computing device comprising:
a data store configured to store media;
a processor communicatively coupled to the data store, the processor configured to:
maintain a model indicating which media file segments are likely to be more interesting to media viewers, the model being trained based on user interactions with individual segments of a training set of media files and interest scores assigned by users to the individual segments of the training set of media files, wherein the user interactions comprise a change in viewership during a presentation of one of the individual segments;
apply, by the computing device, the model to segments of a current media file to generate interest ratings for the segments, wherein an interest rating for a segment is based on a change in viewership during a presentation of the segment;
generate, based on the interest ratings, an edited media file for the current media file, the edited media file including segments of the current media file having interest ratings that meet a threshold criterion; and
add the edited media file to the data store as a summarized version of the current media file.

2. The at least one computing device of claim 1 , wherein the user interactions comprise one or more of viewership repeating a segment, viewership skipping a segment, or viewership pausing on a segment.

3. The at least one computing device of claim 1 , further comprising editing the media file by filtering from the media file those segments of the media file whose interest score falls below the threshold criterion.

4. The at least one computing device of claim 1 , wherein the user interactions comprise one or more of viewership drop-off, viewership full-screen, viewership annotating, viewership sharing, viewership comment posting, viewership chatting, viewership transitioning from a seek action to a play action, or viewership transitioning between a full screen and a small screen.

5. The at least one computing device of claim 1 , further comprising training the model based on a regression analysis.

6. The at least one computing device of claim 1 , wherein the processor is further configured to monitor interactions of a set of viewers with the segments of the current media file, the monitored interactions to be mapped to the interest ratings for the segments of the current media file using the model, the interactions comprising at least one of repeating a segment, skipping a segment, or pausing on a segment.

7. The at least one computing device of claim 1 , wherein the processor is further configured to train the model by:
receiving interest scores assigned to the individual segments of the training set of media files;
measuring viewer metrics for the individual segments of the training set of media files based on the user interactions with the individual segments of the training set of media files; and
forming a set of feature vectors based on the viewer metrics.

8. The at least one computing device of claim 1 , wherein the media file is a video clip.

9. A method implemented by one or more computing devices, the method comprising:
maintaining a model indicating which media file segments are likely to be more interesting to media viewers, the model being trained based on user interactions with individual segments of a training set of media files and interest scores assigned by users to the individual segments of the training set of media files, wherein the user interactions comprise a change in viewership during a presentation of one of the individual segments;
applying, by a processing device, the model to segments of a current media file to generate interest ratings for the segments, wherein an interest rating for a segment is based on a change in viewership during a presentation of the segment;
generating, based on the interest ratings, an edited media file for the current media file, the edited media file including segments of the current media file having interest ratings that meet a threshold criterion; and
storing the edited media file in a data store as a summarized version of the current media file.

10. The method of claim 9 , wherein the user interactions comprise one or more of viewership repeating a segment, viewership skipping a segment, or viewership pausing on a segment.

11. The method of claim 9 , further comprising editing the current media file by filtering from the current media file segments of the media file whose interest score falls below the threshold criterion.

12. The method of claim 9 , further comprising training the model based on regression analysis.

13. The method of claim 9 , wherein applying the model to segments of the current media file comprises monitoring interactions of a set of viewers with the segments of the current media file, the monitored interactions to be mapped to the interest ratings for the segments of the current media file using the model.

14. The method of claim 9 , further comprising training the model by:
receiving interest scores assigned to the individual segments of the training set of media files;
measuring viewer metrics for the individual segments of the training set of media files based on the user interactions with the individual segments of the training set of media files; and
forming a set of feature vectors based on the viewer metrics.

15. One or more non-transitory computer-readable media storing processor-executable instructions that, when executed, cause one or more processors to perform operations comprising:
maintaining a model indicating which media file segments are likely to be more interesting to media viewers, the model being trained based on user interactions with individual segments of a training set of media files and interest scores assigned by users to the individual segments of the training set of media files, wherein the user interactions comprise a change in viewership during a presentation of one of the individual segments;
applying, by the one or more processors, the model to segments of a current media file to generate interest ratings for the segments, wherein an interest rating for a segment is based on a change in viewership during a presentation of the segment;
generating, based on the interest ratings, an edited media file for the current media file, the edited media file including segments of the current media file having interest ratings that meet a threshold criterion; and
storing the edited media file in a data store as a summarized version of the current media file.

16. The one or more non-transitory computer-readable media storing processor-executable instructions of claim 15 , wherein the user interactions comprise one or more of viewership repeating a segment, viewership skipping a segment, or viewership pausing on a segment.

17. The one or more non-transitory computer-readable media storing processor-executable instructions of claim 15 , further comprising editing the current media file by filtering from the current media file segments of the media file whose interest score falls below a threshold.

18. The one or more non-transitory computer-readable media storing processor-executable instructions of claim 15 , further comprising training the model.

19. The one or more non-transitory computer-readable media storing processor-executable instructions of claim 15 , wherein applying the model to segments of the current media file comprises monitoring interactions of a set of viewers with the segments of the current media file, the monitored interactions to be mapped to the i

Assignees

• Google Inc
• Google Llc

Inventors

• Shynar Michael
• Zomet Asaf
• Chechik Gal

Classifications

• G11B27/031 Primary

  Electronic editing of digitised analogue information signals, e.g. audio or video signals · CPC title

• H04N9/8205

  involving the multiplexing of an additional signal and the colour video signal · CPC title

• H04N9/87

  Regeneration of colour television signals (H04N9/80 takes precedence) · CPC title

• H04N5/765

  Interface circuits between an apparatus for recording and another apparatus · CPC title

• H04N21/8549

  Creating video summaries, e.g. movie trailer {(retrieval in video databases by using presentations in form of a video summary G06F16/739)} · CPC title