Method And Device For Tracking Sports Players with Context-Conditioned Motion Models

Abstract

A method and device generates a trajectory. The method includes receiving a plurality of tracklets indicative of movement of a plurality of targets over a predetermined temporal interval. The method includes determining a plurality of context data for a pair of tracklets based upon at least one additional tracklet. The method includes computing a probability that the pair of tracklets relate to a first one of the targets. The method includes generating a trajectory for the first target based upon a concatenation of select ones of the tracklets. The concatenation maximizes the probability that the pair of tracklets correspond to the first target based upon the context data associated with the pair of the tracklets.

1 - 20 . (canceled) 
     
     
         21 . A method, comprising:
 receiving a plurality of tracklets indicative of movement of a plurality of targets over a predetermined temporal interval;   determining context data for a pair of tracklets of the plurality of tracklets based on at least one additional tracklet of the plurality of tracklets;   computing a probability that the pair of tracklets relate to a first target of the plurality of targets, wherein the probability is computed based on a regression; and   generating a trajectory for the first target based on a concatenation of select ones of the plurality of tracklets,   wherein the concatenation maximizes the probability that the pair of tracklets correspond to the first target based on the context data of the pair of tracklets.   
     
     
         22 . The method of claim  1 , wherein the trajectory is used to control an orientation of a measurement device capturing movement of the plurality of targets over the predetermined temporal interval, wherein the orientation corresponds to further movement of the first target of the plurality of targets over a further predetermined time interval. 
     
     
         23 . The method of claim  1 , wherein the measurement device is a camera and the trajectory is used to insert a graphic into a video stream capturing the movement of the plurality of targets. 
     
     
         24 . The method of claim  1 , wherein the measurement device is a camera and the movement of the plurality of targets is captured in a plurality of frames, wherein the plurality of frames each correspond to a different moment occurring within the predetermined temporal interval. 
     
     
         25 . The method of claim  1 , wherein the received plurality of tracklets are determined by:
 detecting responses of the plurality of targets within a local spatial-temporal volume;   identifying clusters of select targets within the local spatial-temporal volume; and   determining random sample consensuses of the select targets within each cluster using constant velocity models.   
     
     
         26 . The method of claim  1 , wherein the plurality of targets are vehicles and the context data indicates a traffic light state. 
     
     
         27 . The method of claim  1 , wherein the plurality of targets are vehicles and the context data indicates to a presence of an accident. 
     
     
         28 . The method of claim  1 , wherein the plurality of targets are part of a crowd of people and the context data indicates anticipated collisions between the plurality of targets. 
     
     
         29 . The method of claim  1 , wherein the context data includes at least one of an absolute occupancy map, a relative occupancy map for a first partial area of the absolute occupancy map, a focus area for a second partial area of the absolute occupancy map in which the second partial area is a subset of the first partial area, and a close interaction detection for the first target and a second target of the plurality of targets. 
     
     
         30 . The method of claim  1 , wherein the regression is a random forest including a plurality of decision trees that is trained to learn a mapping between the pair of tracklets. 
     
     
         31 . A device, comprising:
 a processor coupled to a memory, wherein the processor is programmed to generate a trajectory for a first target of a plurality of targets by:
 receiving a plurality of tracklets indicative of movement of the plurality of targets over a predetermined temporal interval; 
 determining context data for a pair of tracklets of the plurality of tracklets based on at least one additional tracklet of the plurality of tracklets; 
 computing a probability that the pair of tracklets relate to the first target, wherein the probability is computed based on a regression; and 
 generating a trajectory for the first target based on a concatenation of select ones of the plurality of tracklets, 
 wherein the concatenation maximizes the probability that the pair of tracklets correspond to the first target based on the context data of the pair of tracklets. 
   
     
     
         32 . The device of claim  11 , wherein the plurality of tracklets are derived based on a measurement device capturing movement of the plurality of targets over the predetermined temporal interval. 
     
     
         33 . The device of claim  11 , wherein the probability is indicative of whether one of the pair of tracklets is an immediate successor in time to a corresponding tracklet of the pair of tracklets. 
     
     
         34 . The device of claim  11 , wherein the context data is based on a location of each of the plurality of targets relative to a field of a team sports match. 
     
     
         35 . The device of claim  11 , wherein the processor is configured to determine the received plurality of tracklets by:
 detecting responses of the plurality of targets within a local spatial-temporal volume;   identifying clusters of select targets within the local spatial-temporal volume; and   determining random sample consensuses of the select targets within each cluster using constant velocity models.   
     
     
         36 . The device of claim  11 , wherein the plurality of tracklets are based on a set of detections of each of the plurality targets over the predetermined temporal interval; 
     
     
         37 . The device of claim  16 , wherein the plurality of tracklets include incremental tracklets that are determined based on the detection, the incremental tracklets being indicative of an estimate of an instantaneous position and velocity of a respective target. 
     
     
         38 . The device of claim  11 , wherein the regression is a random forest including a plurality of decision trees. 
     
     
         39 . The device of claim  11 , wherein the processor is further configured to:
 generate a further trajectory for a second target of the plurality of targets over the predetermined temporal interval, the further trajectory being a further concatenation of the select ones of the plurality of tracklets,   wherein the further concatenation maximizes a probability that a further one of the tracklets corresponds to the second target based on the context data associated with the further pair of tracklets.   
     
     
         40 . A non-transitory computer readable storage medium with an executable program stored thereon, wherein the program instructs a processor to perform operations comprising:
 receiving a plurality of tracklets indicative of movement of a plurality of targets over a predetermined temporal interval;   determining context data for a pair of tracklets of the plurality of tracklets based on at least one additional tracklet of the plurality of tracklets;   computing a probability that the pair of tracklets relate to a first target of the plurality of targets, wherein the probability is computed based on a regression; and   generating a trajectory for the first target based on a concatenation of select ones of the tracklets,   wherein the concatenation maximizes the probability that the pair of tracklets correspond to the first target based on the context data of pair of the tracklets.