Tracking player role using non-rigid formation priors

Patent metadata

Field	Value
Publication number	US-9342785-B2
Application number	US-201314081454-A
Country	US
Kind code	B2
Filing date	Nov 15, 2013
Priority date	Nov 15, 2013
Publication date	May 17, 2016
Grant date	May 17, 2016

Abstract

Approaches are described for assigning roles to agents in a group of agents engaging in an activity. An assignment analysis system receives a first set of detections, where each detection in the first set of detections comprises a physical location. The assignment analysis system defines an exemplar formation comprising an arrangement of each role in a set of roles. The assignment analysis system calculates a first cost function between at least one detection in the first set of detections and at least one role in the set of roles. The assignment analysis system generates a first set of permutations based on the first cost function. The assignment analysis system assigns a first role in the set of roles to a first detection in the first set of detections based on the first set of permutations.

First claim

What is claimed is:

1. A method for assigning roles to agents in a first group of agents engaging in an activity, comprising:
receiving a first set of detections, wherein each detection in the first set of detections comprises a physical location;
defining an exemplar formation comprising an arrangement of each role in a set of roles;
calculating a first cost function between at least one detection in the first set of detections and at least one role in the set of roles;
generating a first set of permutations based on the first cost function; and
assigning a first role in the set of roles to a first detection in the first set of detections based on the first set of permutations.

2. The method of claim 1 , further comprising separating the first set of detections into a first portion comprising physical locations associated with the first group of agents and a second portion comprising physical locations associated with a second group of agents.

3. The method of claim 1 , wherein each physical location in the first set of physical locations comprises an absolute location in a playing field configured for a team sport.

4. The method of claim 1 , wherein each physical location in the first set of physical locations comprises a position of a first agent in the first group of agents relative to a second agent in the first group of agents.

5. The method of claim 1 , wherein assigning a first role in the set of roles to a first detection in the first set of detections comprises:
selecting a set of potential permutations for evaluation, wherein each permutation defines a relationship between each detection in the set of detections and each role on the set of roles;
for each time step in a set of time steps, calculating a transition cost to transition from a first permutation at a first time step to a second permutation at a second time step;
selecting a lowest cost path through the set of time steps based on the transition costs; and
assigning the first role in the set of roles to the first detection in the first set of detections based on the lowest cost path.

6. The method of claim 1 , wherein assigning a first role in the set of roles to a first detection in the first set of detections comprises:
selecting a set of potential permutations for evaluation, wherein each permutation defines a relationship between each detection in the set of detections and each role on the set of roles;
for each time step in a set of time steps, calculating a transition cost to transition from a first permutation at a first time step to a second permutation at a second time step;
selecting a lowest cost path through the set of time steps based on the transition costs;
determining that the lowest cost path is temporally inconsistent with a current assignment of roles in the set of roles to detections in a second set of detections;
selecting a next lowest cost path through the set of time steps based on the transition costs; and
assigning the first role in the set of roles to the first detection in the first set of detections based on the next lowest cost path.

7. The method of claim 1 , further comprising:
calculating a second cost function between at least one detection in the first set of detections and at least one agent in the first group of agents;
generating a second set of permutations based on the second cost function; and
assigning a first identity associated with a first agent in the first group of agents to a first detection in the first set of detections based on the second set of permutations.

8. The method of claim 7 , further comprising:
receiving a second set of detections, wherein each detection in the second set of detections comprises a physical location;
calculating a third cost function between at least one detection in the second set of detections and at least one role in the set of roles;
generating a third set of permutations based on the third cost function;
calculating a fourth cost function between at least one detection in the second set of detections and at least one agent in the first group of agents; and
generating a fourth set of permutations based on the fourth cost function;
wherein assigning the first role to the first detection is further based on at least one of the second set of permutations, the third set of permutations, and the fourth set of permutations, and wherein assigning the first identity to the first detection is further based on at least one of the first set of permutations, the third set of permutations, and the fourth set of permutations.

9. The method of claim 1 , further comprising:
determining that a first quantity of detections in the first set of detections is less than a second quantity of roles in the set of roles; and
selecting a second role in the set of roles that lacks a corresponding detection in the first set of detections based on the first cost function.

10. The method of claim 1 , further comprising:
determining that a first quantity of detections in the first set of detections is greater than a second quantity of roles in the set of roles; and
selecting a second detection in the first set of detections that lacks a corresponding role in the set of roles based on the first cost function.

11. The method of claim 1 , wherein the exemplar formation is selected by a user from a predefined set of exemplar formations.

12. The method of claim 1 , further comprising characterizing a behavior of the first group of agents based at least in part on the assignment of the first role to the first detection, wherein the behavior is defined by a plurality of sets of detections including the first set of detections.

13. The method of claim 1 , further comprising:
retrieving a first sequence of detections based at least in part on the assignment of the first role to the first detection, wherein the first sequence of detections comprises a plurality of sets of detections including the first set of detections, each set of detections corresponding to a different time;
retrieving an input feature representing a characteristic of the first sequence of detections;
identifying a second sequence of detections included in a database of detection data based on the input feature; and
retrieving the second sequence of detections from the database.

14. A non-transitory computer-readable storage medium including instructions that, when executed by a processor, cause the processor to assign roles to agents in a first group of one or more agents engaging in an activity, by performing an operation comprising:
receiving a first set of detections, wherein each detection in the first set of detections comprises a physical location;
defining an exemplar formation comprising an arrangement of each role in a set of roles;
calculating a first cost function between at least one detection in the first set of detections and at least one role in the set of roles;
generating a first set of permutations based on the first cost function; and
assigning a first role in the set of roles to a first detection in the first set of detections based on the first set of permutations.

15. The non-transitory computer-readable storage medium of claim 14 , further comprising separating the first set of detections into a first portion comprising physical locations associated with the first group of agents and a second portion comprising physical locations associated with a second group of agents.

16. The non-transitory computer-readable storage medium of claim 14 , wherein each physical location in the first set of physical locations comprises an absolute location in a

Assignees

• Disney Entpr Inc

Inventors

• Lucey Patrick
• Bialkowski Alina
• Carr G Peter
• Matthews Iain
• Sheikh Yaser

Classifications

• G06N5/022 Primary

  Knowledge engineering; Knowledge acquisition · CPC title

• G06K9/00624

  Physics · mapped topic

• G06K9/00724

  Physics · mapped topic

• G06V20/42

  of sport video content · CPC title

• G06V20/00

  Scenes; Scene-specific elements (control of digital cameras H04N23/60) · CPC title