Methods and systems of spatiotemporal pattern recognition for video content development

• US 10,748,008 B2
• Filed: 03/12/2019
• Issued: 08/18/2020
• Est. Priority Date: 02/28/2014
• Status: Active Grant

• Current Assignee: Genius Sports SS LLC
• Original Assignee: Second Spectrum, Inc. (Genius Sports Ltd.)
• Inventors: Chang, Yu-Han, Maheswaran, Rajiv Tharmeswaran, Su, Jeffrey Wayne, Hollingsworth, Noel
• Primary Examiner(s): Tran, Loi H
• Application Number:
  US16/351,213
• Publication Number:
  US 20190205651A1
• Time in Patent Office: 525 Days
• Field of Search: 725 19
• CPC Class Codes: A63F 13/60 Generating or modifying game content before or while executing the game program, e.g. authoring tools specially adapted for game development or game-integrated level editor
• CPC Class Codes: G06F 3/012 Head tracking input arrangements
• CPC Class Codes: G06F 3/013 Eye tracking input arrangements G06F3/015 takes precedence
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30221 Sports video; Sports image
• CPC Class Codes: G06V 20/42 of sport video content
• CPC Class Codes: G06V 20/46 Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
• CPC Class Codes: G11B 27/031 Electronic editing of digitised analogue information signals, e.g. audio or video signals
• CPC Class Codes: G11B 27/28 by using information signals recorded by the same method as the main recording G11B27/22 takes precedence
• CPC Class Codes: H04N 13/117 the virtual viewpoint locations being selected by the viewers or determined by viewer tracking
• CPC Class Codes: H04N 13/204 using stereoscopic image cameras stereoscopic photography G03B35/00
• CPC Class Codes: H04N 13/243 using three or more 2D image sensors
• CPC Class Codes: H04N 21/2187 Live feed
• 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 21/251 Learning process for intelligent management, e.g. learning user preferences for recommending movies details of learning user preferences for the retrieval of video data in a video database G06F16/739; computer systems using learning methods G06N3/08
• CPC Class Codes: H04N 21/4223 Cameras H04N23/00 takes precedence
• CPC Class Codes: H04N 21/4345 Extraction or processing of SI, e.g. extracting service information from an MPEG stream
• CPC Class Codes: H04N 21/44008 involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream arrangements characterised by components specially adapted for monitoring, identification or recognition of video in broadcast systems H04H60/59
• CPC Class Codes: H04N 21/4532 : involving end-user characteristics, e.g. viewer profile, preferences monitoring of user activities for profile generation for accessing a video database G06F16/739; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18
• CPC Class Codes: H04N 21/4662 : characterized by learning algorithms
• CPC Class Codes: H04N 21/8456 : by decomposing the content in the time domain, e.g. in time segments
• CPC Class Codes: H04N 21/8549 : Creating video summaries, e.g. movie trailer retrieval in video databases by using presentations in form of a video summary G06F16/739
• CPC Class Codes: H04N 5/2224 : related to virtual studio applications

Abstract

Providing enhanced video content includes processing at least one video feed through at least one spatiotemporal pattern recognition algorithm that uses machine learning to develop an understanding of a plurality of events and to determine at least one event type for each of the plurality of events. The event type includes an entry in a relationship library detailing a relationship between two visible features. Extracting and indexing a plurality of video cuts from the video feed is performed based on the at least one event type determined by the understanding that corresponds to an event in the plurality of events detectable in the video cuts. Lastly, automatically and under computer control, an enhanced video content data structure is generated using the extracted plurality of video cuts based on the indexing of the extracted plurality of video cuts.

Claims

1. A method for providing enhanced video content, comprising:

processing at least one video feed through at least one spatiotemporal pattern recognition algorithm that uses a machine learning system to determine at least one event type for each of a plurality of events within the at least one video feed;

extracting a plurality of video cuts from the at least one video feed using a combination of the determination of the at least one event type and a determination of an event type of another input feed selected from the group consisting of a broadcast video feed, an audio feed and a closed caption feed, wherein the determination of an event type of the another input feed is responsive to applying the machine learning system to a portion of content of a broadcast commentary present in the another input feed;

indexing the extracted plurality of video cuts based on a similarity of the at least one event type determined by the machine learning system with the event type determined by applying the machine learning system to the portion of content of broadcast commentary present in the another input feed;

identifying at least one pattern in the extracted plurality of video cuts, the pattern spanning a plurality of frames of at least one of the extracted plurality of video cuts; and

indexing at least a portion of the plurality of the extracted video cuts that comprise the at least one pattern with an attribute indicative of the at least one pattern;

wherein the at least one spatiotemporal pattern recognition algorithm is based on at least one pattern recognized by adjusting an input feature of a plurality of input features and a weight thereof within the machine learning system, wherein the machine learning system is constructed to process the plurality of input features of the at least one video feed, the plurality of input features comprising;

relative direction of motion of at least two visible features,

duration of relative motion of visible features with respect to each other,

rate of motion of at least two visible features with respect to each other,

relative acceleration of motion of at least two visible features, and

relative projected point of intersection of at least two visible features.

2. The method of claim 1, wherein the at least one pattern is determined by applying the machine learning system to the extracted plurality of video cuts.

3. The method of claim 2, wherein the determining by the machine learning system includes identifying at least one player involved in an event, and wherein indexing the extracted plurality of video cuts includes identifying at least one player represented in at least one of the video cuts from the plurality of the video cuts.

4. The method of claim 2, wherein the at least one pattern comprises a series of same event types involving a same player over time.

5. The method of claim 1, wherein the plurality of video cuts includes a player during multiple, identical event types over time.

6. The method of claim 1, further comprising providing an enhanced video feed that shows a player during the plurality of events over time, wherein the enhanced video feed is at least one of a simultaneous, superimposed video of the player involved in multiple, identical event types and a sequential video of the player involved in an event type.

7. The method of claim 1, wherein identifying the at least one pattern includes identifying sequences of events that predict a given action that is likely to follow.

8. The method of claim 1, wherein identifying the at least one pattern includes identifying similar sequences of events across the plurality of video cuts.

9. The method of claim 1, further comprising providing a user interface that enables a user to at least one of view and interact with the at least one pattern.

10. The method of claim 9, wherein the at least one pattern and at least one interaction option are personalized based on at least one of a user preference and a user profile.

11. The method of claim 1, wherein the at least one pattern relates to an anticipated outcome of at least one of a game and an event within a game.

12. The method of claim 11, further comprising providing a user with at least one of a statistic, trend information and a prediction based on the at least one pattern.

13. The method of claim 12, wherein the at least one of the statistic, the trend information or the prediction is based on at least one of a user preference and a user profile.

14. The method of claim 1, wherein the at least one pattern relates to play of an athlete.

15. The method of claim 14, further comprising providing a comparison of the play of the athlete with another athlete based on a similarity of at least one of a portion of the extracted plurality of video cuts for the athlete and the another athlete and the at least one pattern identified for the athlete and the another athlete.

16. The method of claim 15, wherein the comparison is between the play of a professional athlete and the play of a non-professional user.

17. The method of claim 16, wherein the comparison is based on a similarity of a playing style of the professional athlete, as determined by the machine learning system of at least one event of the plurality of events and the at least one pattern, with at least one feature of the playing style of the non-professional user.

18. The method of claim 1, wherein the determining the at least one event type for each of the plurality of events further comprises using the plurality of events in position tracking data over time obtained from at least one of the at least one video feed and a chip-based player tracking system, and wherein the determining the at least one event type of each of the plurality of events is based on at least two of spatial configuration, relative motion, and projected motion of at least one of a player and an item used in a game.

19. The method of claim 1, wherein the determining the at least one event type of each of the plurality of events further comprises aligning multiple unsynchronized input feeds related to an event of the plurality of events using at least one of a hierarchy of algorithms and a hierarchy of human operators, wherein the unsynchronized input feeds are selected from the group consisting of one or more broadcast video feeds of the event, one or more feeds of tracking video for the event, and one or more play-by-play data feeds of the event.

20. The method of claim 19, wherein the multiple unsynchronized input feeds include at least three feeds selected from at least two types related to the event.

21. The method of claim 19, further comprising at least one of validating and modifying the alignment of the unsynchronized input feeds using a hierarchy involving at least two of one or more algorithms, one or more human operators, and one or more input feeds.

22. The method of claim 1, further comprising at least one of validating the determination and modifying the determination of the at least one event type using a hierarchy involving at least two of one or more algorithms, one or more human operators, and one or more input feeds.