Dynamic classification system for sports analysis

• US 10,019,630 B1
• Filed: 01/09/2017
• Issued: 07/10/2018
• Est. Priority Date: 01/09/2017
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Poletin, Lea, Rickayzen, Maximilian
• Primary Examiner(s): Kassa, Yosef
• Application Number:
  US15/401,566
• Publication Number:
  US 20180197013A1
• Time in Patent Office: 547 Days
• Field of Search: 382190, 382201, 382209, 382224, 382226, 382228, 473493, 473479, 473416
• CPC Class Codes: G06V 20/42 of sport video content

Abstract

Provided are devices and methods for rating video scenes from a sporting event. In one example, the method includes receiving data of a plurality of rated sports scenes, normalizing the data of the plurality of rated sports scenes into a plurality of data points, and forming at least two rating groups of data points in dimensional space, normalizing an unrated sports scene into a data point, determining a rating for the unrated scene by comparing the data point of the unrated scene to the at least two rating groups of data points, and outputting information about the determination to a display of a user device. According to various embodiments, the device and method may learn from a few plays manually rated by a coach and apply the coach'"'"'s learned philosophy automatically to a plurality of unrated plays.

Claims

1. A computing device for rating scenes of video from a sporting event, the computing device comprising:

a storage configured to store data of a plurality of rated sports scenes;

a processor configured to normalize the data of the plurality of rated sports scenes into a plurality of data points in dimensional space, form at least two rating groups based on the plurality of data points in the dimensional space, normalize an unrated sports scene into a data point, and determine a rating for the unrated scene by comparing the data point of the unrated scene to the at least two rating groups in the dimensional space; and

an output configured to output information about the determined rating for the unrated sports scene to a display of a user device.

2. The computing device of claim 1, wherein the data of the plurality of sports scenes comprises, for each scene, Extensible Markup Language (XML) data.

3. The computing device of claim 1, wherein the processor is configured to normalize the data of each rated sports scene based on a plurality of predetermined features of the scene to generate a data point in a multidimensional feature space in which each predetermined feature represents a dimension in the feature space.

4. The computing device of claim 3, wherein the processor is further configured to assign weights to each of the predetermined features by assigning weights to the dimensions in the multidimensional feature space.

5. The computing device of claim 1, wherein the processor is further configured to identify a first cluster of data points in the dimensional space corresponding to relevant scenes and a second cluster of data points in the dimensional space corresponding to non-relevant scenes, and determine a representative data point of a relevant scene from the first cluster and a representative data point of a non-relevant scene from the second cluster.

6. The computing device of claim 5, wherein the processor is further configured to determine a center of the first cluster of data points in the dimensional space as the representative data point of a relevant scene, and determine a center of the second cluster of data points in the dimensional space as the representative data point of a non-relevant scene.

7. The computing device of claim 5, wherein the processor is further configured to determine whether the unrated sports scene is a relevant scene or a non-relevant scene based on a distance between the data point corresponding to the unrated scene and the representative data point of the relevant scene and a distance between the data point corresponding to the unrated scene and the representative data point of the non-relevant scene.

8. The computing device of claim 1, wherein the output is configured to output a table comprising a row representing the unrated sports scene, and a plurality of columns representing a plurality of predetermined features of the unrated sports scene including the determined rating.

9. A method for rating scenes of video from a sporting event, the method comprising:

receiving data of a plurality of rated sports scenes;

normalizing the data of the plurality of rated sports scenes into a plurality of data points in dimensional space, and forming at least two rating groups based on the plurality of data points in the dimensional space;

normalizing an unrated sports scene into a data point;

determining a rating for the unrated scene by comparing the data point of the unrated scene to the at least two rating groups of data points in the dimensional space; and

outputting information about the determined rating for the unrated sports scene to a display of a user device.

10. The method of claim 9, wherein the data of the plurality of sports scenes comprises, for each scene, Extensible Markup Language (XML) data.

11. The method of claim 9, wherein the normalizing the data of each rated sports scene is performed based on a plurality of predetermined features of the scene to generate a data point in a multidimensional feature space in which each predetermined feature represents a dimension in the feature space.

12. The method of claim 11, further comprising assigning weights to each of the predetermined features by assigning weights to the dimensions in the multidimensional feature space.

13. The method of claim 9, wherein the identifying the at least two rating groups comprises identifying a first cluster of data points in the dimensional space corresponding to relevant scenes and a second cluster of data points in the dimensional space corresponding to non-relevant scenes, and determining a representative data point of a relevant scene from the first cluster and a representative data point of a non-relevant scene from the second cluster.

14. The method of claim 13, wherein the identifying further comprises determining a center of the first cluster of data points in the dimensional space as the representative data point of a relevant scene, and determining a center of the second cluster of data points in the dimensional space as the representative data point of a non-relevant scene.

15. The method of claim 13, wherein the determining comprises determining whether the unrated sports scene is a relevant scene or a non-relevant scene based on a distance between the data point corresponding to the unrated scene and the representative data point of the relevant scene and a distance between the data point corresponding to the unrated scene and the representative data point of the non-relevant scene.

16. The method of claim 9, wherein the outputting comprises outputting a table comprising a row representing the unrated sports scene, and a plurality of columns representing a plurality of predetermined features of the unrated sports scene including the determined rating.

17. A non-transitory computer readable medium having stored therein instructions that when executed cause a computer to perform a method for rating scenes of video from a sporting event, the method comprising:

receiving data of a plurality of rated sports scenes;

normalizing the data of the plurality of rated sports scenes into a plurality of data points in dimensional space, and forming at least two rating groups based on the plurality of data points in the dimensional space;

normalizing an unrated sports scene into a data point;

determining a rating for the unrated scene by comparing the data point of the unrated scene to the at least two rating groups of data points in the dimensional space; and

outputting information about the determined rating for the unrated sports scene to a display of a user device.

18. The computer-readable medium of claim 17, wherein the data of the plurality of sports scenes comprises, for each scene, Extensible Markup Language (XML) data.

19. The computer-readable medium of claim 17, wherein the normalizing the data of each rated sports scene is performed based on a plurality of predetermined features of the scene to generate a data point in a multidimensional feature space in which each predetermined feature represents a dimension in the feature space.

20. The computer-readable medium of claim 17, wherein the outputting comprises outputting a table comprising a row representing the unrated sports scene, and a plurality of columns representing a plurality of predetermined features of the unrated sports scene including the determined rating.