Method, apparatus and computer program for generating sports game highlight video based on excitement of gameplay
View Patent ↗A method for generating a sports game highlight video based on excitement of gameplay includes receiving video of the sports game from at least one image capture device; performing object tracking analysis on the video to generate tracking data providing object trajectory and player position information; identifying shot events in the video of the sports game based on said tracking data; extracting video segments of the identified shot events from the video of the sports game; generating shot event metadata for indexing the video segments; calculating a point excitement matrix by a point excitement model based on said shot event metadata; and generating a sports game highlight video based on said point excitement matrix.
1 . A method for generating a sports game highlight video based on excitement of gameplay, comprising the steps of:
receiving video of the sports game from at least one image capture device;
performing object tracking analysis on the video to generate tracking data providing object trajectory and player position information;
identifying shot events in the video of the sports game based on said tracking data, wherein the shot events represent respective series of events for respective shots;
extracting video segments of the identified shot events from the video of the sports game, wherein the video segments extracted include the respective series of events for the respective shots;
generating shot event metadata for the shot events identified and associating the shot event metadata with the video segments extracted;
calculating a point excitement matrix for a game element of the sports game, by a point excitement model, based on said shot event metadata, wherein the point excitement matrix includes a set of game excitement weights, wherein game excitement weights of the set of game excitement weights represent an excitement of the sports game with respective to the game element; and
generating a sports game highlight video based on said point excitement matrix.
2 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a shot based on the occurrence of a shot category and likelihood of a winner shot in a rally.
3 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating of the point excitement matrix comprises evaluating a game excitement weight with respect to a rally based on rally length, and wherein the rally length is measured by the number of shots in the rally or the absolute time taken by the rally.
4 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a winner and an error by the player based on frequency of winner, or frequency of forced errors or frequency of unforced errors.
5 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a rally based on the quality of the final shot in a rally.
6 . The method for generating a sports game highlight video according to claim 5 , wherein the quality of the final shot in a rally is calculated by multiplying unreturnability of the final shot in a rally with a game excitement weight with respect to a winner and an error by the player based on frequency of winner, or frequency of forced errors or frequency of unforced errors.
7 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a sports game based on overall shot quality in the sports game.
8 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a return shot based on return improbability.
9 . The method for generating a sports game highlight video according to claim 8 , wherein the return improbability is calculated by multiplying incoming shot unreturnability with return shot unreturnability.
10 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to a rally based on average rally unreturnability across all shots within the rally.
11 . The method for generating a sports game highlight video according to claim 6 , wherein the unreturnability of a shot is calculated based on the position of player receiving the shot, trajectory of the shot, speed of the shot, forward or backward stroke, or historical data of player performance.
12 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises evaluating a game excitement weight with respect to point state information, wherein the point state information is obtained from a scoring feed or by analysing the video of the sports game.
13 . The method for generating a sports game highlight video according to claim 1 , wherein the calculating of the point excitement matrix is further based on excitement flags added to the video of the sports game by a user marking exciting moments in the sports game.
14 . The method for generating a sports game highlight video according to claim 1 , wherein the calculating of the point excitement matrix is further based on at least one of analysis of commentator reaction in the video of the sports game, analysis of spectator reaction in the video of the sports game, or further based on analysis of an audio track in the video of the sports game, wherein the audio track includes at least one of commentator audio, spectator audio, or player audio.
15 . The method for generating a sports game highlight video according to claim 1 , wherein the step of calculating the point excitement matrix comprises using a machine learning algorithm trained by historical data including shot event metadata.
16 . The method for generating a sports game highlight video according to claim 1 , wherein the step of performing video editing further comprises:
filtering and ranking the video segments according to at least one game excitement weight in said point excitement matrix; and
generating a sports game highlight video by concatenating the filtered and ranked video segments.
17 . The method for generating a sports game highlight video according to claim 16 , wherein the step of performing video editing further comprises:
generating tracking model videos from tracking data; and
merging video segments with tracking model videos corresponding to the same shot.
18 . The method for generating a sports game highlight video according to claim 1 , wherein the sports game is selected from the group consisting of tennis, badminton, table tennis, squash and volleyball.
19 . A non-transitory computer readable medium storing a computer readable instructions that, when loaded onto a computer, configure the computer to perform the method according to claim 1 .
20 . The method of claim 1 , wherein the shot event metadata for a shot event includes one or more of point state information, shot type, rally length, winner information, error information, shot quality, unreturnability information, or return improbability.