Apparatus and method for image classification
View Patent ↗An image classification system for tracking a net sport is illustrated. The system includes a visual device configured to capture video footage of a match. The computing device is communicatively connected to the visual device and is configured to receive the video footage of the match from the visual device, translate video footage of the match from the visual device into a data representation by logging an event with a timestamp, and display data representation to a user.
1. An apparatus for image classification of moving objects, the apparatus comprising:
a processor; and
a memory communicatively connected to the processor, the memory containing instructions configuring the processor to:
receive a net sport match video footage of a plurality of moving objects from a visual device, wherein the plurality of moving objects comprises at least a ball and one or more players;
detect an event in the video footage, wherein the event comprises a collision;
determine the one or more players of the plurality of moving objects; and
display a data representation of the event to a user, wherein displaying comprises:
receiving a plurality of temporal sections of the net sport match video footage related to the event and the one or more players;
generating a highlights score for each temporal section of the plurality of temporal sections;
ranking the plurality of temporal sections based on the highlights score; and
generating auto-highlights video based on the ranked plurality of temporal sections.
2. The apparatus of claim 1 , wherein the visual device comprises a camera.
3. The apparatus of claim 1 , wherein video footage comprises a video taken from at least a back-view.
4. The apparatus of claim 1 , wherein detecting the event comprises comparing a series of frames of the video footage to determine movement of each of the plurality of moving objects.
5. The apparatus of claim 4 , wherein comparing the series of frames comprises video compression by inter-frame coding.
6. The apparatus of claim 1 , wherein determining the one or more players comprises distinguishing humans from other moving objects within the video footage using a player machine-learning model.
7. The apparatus of claim 1 , wherein the collision comprises the ball in the video footage contacting another object.
8. The apparatus of claim 1 , wherein detecting the event further comprises tracking a location of the one or more players.
9. The apparatus of claim 8 , wherein tracking a location of the one or more players comprises determining a location of a player on a court by identifying a player pose skeleton of each of the one or more players.
10. The apparatus of claim 8 , wherein tracking a location of the one or more players comprises human re-identification.
11. The apparatus of claim 1 , wherein the data representation comprises statistics created by data points of the video footage.
12. The apparatus of claim 11 , wherein the data points comprise a velocity of the ball.
13. The apparatus of claim 11 , wherein the data points comprise identifiers of the one or more players.
14. The apparatus of claim 11 , wherein the data points comprise an action path of the one or more players.
15. A method for image classification of moving objects, the method comprising:
receiving, by a processor, a net sport match video footage of a plurality of moving objects from a visual device, wherein the plurality of moving objects comprises at least a ball and one or more players;
detecting, by the processor, an event in the video footage, wherein the event comprises a collision;
determining, by the processor, the one or more players of the plurality of moving objects; and
displaying, by the processor, a data representation of the event to a user, wherein displaying comprises:
receiving a plurality of temporal sections of the net sport match video footage related to the event and the one or more players;
generating a highlights score for each temporal section of the plurality of temporal sections;
ranking the plurality of temporal sections based on the highlights score; and
generating auto-highlights video based on the ranked plurality of temporal sections.
16. The method of claim 15 , wherein the visual device comprises a camera.
17. The method of claim 15 , wherein detecting the event comprises comparing a series of frames of the video footage to determine movement of each of the plurality of moving objects.
18. The method of claim 17 , wherein comparing the series of frames comprises video compression by inter-frame coding.
19. The method of claim 15 , wherein detecting the event further comprises tracking a location of the one or more players.
20. The method of claim 19 , wherein tracking a location of the one or more players comprises determining a location of a player on a court by identifying a player pose skeleton of each of the one or more players.