- Number
- 10062033
- Published
- 2018-08-28
- Filed
- 2014-09-26
- Assignee
- Disney Enterprises, Inc.
- Inventors
- Lucey; Patrick et al.
- CPC
- G06V20/42; G06N7/01
- Verdict
- Set aside sports team-formation analytics/business
- Source
- Google Patents · FreePatentsOnline
Abstract
Approaches are described for discovering a formation associated with an agent group engaging in an activity over a window of time. A formation analysis system computes first and second results for an objective function based on first and second sets of role assignments for each agent in the agent group at first and second moments in time, respectively. The formation analysis system iterates by: replacing the first set of role assignments with the second set of role assignments, and determining whether completion criteria have been met based at least in part on comparing the first result with the second result. If the completion criteria have not been met, then the formation analysis system replaces the second set of role assignments with a third set of role assignments that associate each agent in the first agent group with a different role assignment in the third set of role assignments at a third moment in time. If the completion criteria have been met, then the formation analysis system determines the first formation based on the second result.
Background
BACKGROUND(1) Field(2) The present application relates to artificial intelligence systems used to analyze participants engaged in an activity and, in particular, to analysis of team behaviors using role and formation information.(3) Description of the Related Art(4) In possession-based sports, such as soccer and basketball, players typically maintain a spatial arrangement or “formation,” where a formation is a formal arrangement of a number of persons or things acting as a unit in order to optimize the chance of achieving one or more objectives. The formation is generally defined in terms of a set of roles, where players adopt specific roles associated with the formation. As play progresses, players may change or swap roles to achieve team objectives, such as to respond to the adversarial team's movements or to score a goal. A team may dynamically change from one formation to another, which may involve some or all players adopting a new role associated with the new position.(5) Vision-based player tracking systems have been deployed to track the movement of the players during a game or match. Tracking systems may also track the movement of an object of play, such as the position of the ball in possession-based sports. This tracking data is often archived over multiple games, whereby coaches, sports analysts, and broadcasters may access tracking data for entire games spanning one or more seasons of a given sport. In order to make use of this tracking data, the coach, analyst,
Claims
1. A method comprising: receiving a plurality of data frames specifying location data for each of a plurality of agents, at a plurality of discrete time instances over a window of time, wherein the location data comprises a location of the respective agent in a physical space, and wherein the location data is collected using one or more electronic sensor devices; creating an initial assignment of a plurality of roles to the plurality of agents for the window of time; generating a plurality of probability distribution functions for the plurality of roles, each defining a location of an agent assigned to each respective role during the window of time; iteratively determining, by operation of one or more computer processors, a first formation of the plurality of agents over the window of time, comprising, for each of one or more iterations: creating, for each of the plurality of data frames, an assignment of the plurality of roles to the plurality of agents using the plurality of probability distribution functions; modifying the plurality of probability distribution functions based on the assignment of the plurality of roles to the plurality of agents, wherein each of the modified probability distribution functions defines a location of an agent assigned to each respective role in each respective data frame; retrieving one or more instances of digital video content from a content repository using the first formation, wherein each of the one or more instances of digital video content depicts a formation that is similar to the first formation; and transmitting the one or more instances of digital video content for output to a client.
9. A non-transitory computer-readable storage medium including instructions that, when executed by a processor, cause the processor to perform an operation comprising: receiving a plurality of data frames specifying location data for each of a plurality of agents, at a plurality of discrete time instances over a window of time, wherein the location data comprises a location of the respective agent in a physical space, and wherein the location data is collected using one or more electronic sensor devices; creating an initial assignment of a plurality of roles to the plurality of agents for the window of time; generating a plurality of probability distribution functions for the plurality of roles, each defining a location of an agent assigned to each respective role during the window of time; iteratively determining, by operation of one or more computer processors, a first formation of the plurality of agents over the window of time, comprising, for each of one or more iterations: creating, for each of the plurality of data frames, an assignment of the plurality of roles to the plurality of agents using the plurality of probability distribution functions; modifying the plurality of probability distribution functions based on the assignment of the plurality of roles to the plurality of agents, wherein each of the modified probability distribution functions defines a location of an agent assigned to each respective role in each respective data frame; retrieving one or more instances of digital video content from a content repository using the first formation, wherein each of the one or more instances of digital video content depicts a formation that is similar to the first formation; and transmitting the one or more instances of digital video content for output to a client.
15. A computing system, comprising: a memory that is configured to store instructions for a program; and a processor that is configured to execute the instructions for the program to perform an operation comprising: receiving a plurality of data frames specifying location data for each of a plurality of agents, at a plurality of discrete time instances over a window of time, wherein the location data comprises a location of the respective agent in a physical space, and wherein the location data is collected using one or more electronic sensor devices; creating an initial assignment of a plurality of roles to the plurality of agents for the window of time; generating a plurality of probability distribution functions for the plurality of roles, each defining a location of an agent assigned to each respective role during the window of time; iteratively determining, by operation of one or more computer processors, a first formation of the plurality of agents over the window of time, comprising, for each of one or more iterations: creating, for each of the plurality of data frames, an assignment of the plurality of roles to the plurality of agents using the plurality of probability distribution functions; modifying the plurality of probability distribution functions based on the assignment of the plurality of roles to the plurality of agents, wherein each of the modified probability distribution functions defines a location of an agent assigned to each respective role in each respective data frame; retrieving one or more instances of digital video content from a content repository using the first formation, wherein each of the one or more instances of digital video content depicts a formation that is similar to the first formation; and transmitting the one or more instances of digital video content for output to a client.