System for annotating media content for automatic content understanding
A method to correct for temporal variability in incoming streams of media and data to optimize the performance of a pattern recognition system includes the steps of receiving from one of the incoming streams a point in time when an event is announced, applying probability distribution about the point in time, shifting a point of highest probability of the probability distribution back in time by an amount effective to accommodate for a delay between the event and the announcement, comparing a detected pattern of the event to a stored pattern of similar events and applying a confidence value to the comparison, and confirming to the pattern recognition system that the event occurred at the point of highest probability when the confidence score exceeds a predefined threshold. The method is useful to determine the time at which a particular play occurs during a sporting event, such as the time of a shot-on-goal in a soccer match.
1. A method comprising:
receiving, by a system including a processor, ground truth metadata indicating that a timed activity is visible in a live media stream;
receiving, by the system, a game clock value synchronized to the live media stream;
performing, by the system, optical character recognition on the live media stream according to the game clock value to identify game clock portions of frames of the live media stream;
determining, by the system, variances between sequences of the game clock portions of the frames of the live media stream by computing pixel intensity differences at pixel positions in the game clock portions;
identifying, by the system, static miniboard pixels and dynamic miniboard pixels according to the variances between the sequences of the game clock portions of the frames of the live media stream;
marking, by the system, positions of the static miniboard pixels and the dynamic miniboard pixels of the sequences of the game clock portions of the frames of the live media stream; and
recognizing, by the system, a miniboard display based on the marking of the positions of the static miniboard pixels and the dynamic miniboard pixels without utilizing additional ground truth metadata.
2. The method of claim 1 , wherein the timed activity is a sporting event.
3. The method of claim 1 , wherein the recognizing the miniboard display is in real time with receiving of the live media stream.
4. The method of claim 1 , wherein the marking the positions of the static miniboard pixels is based on marking boundaries of recognized characters of static information portions of the sequences of the game clock portions of the frames of the live media stream.
5. The method of claim 1 , wherein the frames of the live media stream are latest received frames.
6. A device comprising:
a processing system including a processor; and
a memory that stores executable instructions that, when executed by the processing system, facilitate performance of operations, comprising:
determining that a timed activity is visible in a live media stream;
receiving a game clock value synchronized to the live media stream;
performing optical character recognition on the live media stream according to the game clock value to identify game clock portions of frames of the live media stream;
determining variances between sequences of the game clock portions of the frames of the live media stream by computing pixel intensity differences at pixel positions in the game clock portions;
identifying, by the system, static miniboard pixels and dynamic miniboard pixels according to the variances between the sequences of the game clock portions of the frames of the live media stream;
marking positions of the static miniboard pixels and the dynamic miniboard pixels; and
recognizing a miniboard display based on the marking of the positions.
7. The device of claim 6 , wherein the determining that the timed activity is visible in the live media stream is based on metadata associated with the live media stream, and wherein the executable instructions further facilitate the operations of determining that the timed activity is not visible in the live media stream, thereby avoiding a false detection of the miniboard.
8. The device of claim 7 , wherein the recognizing the miniboard display is without receiving and utilizing additional metadata.
9. The device of claim 6 , wherein the timed activity is a sporting event.
10. The device of claim 6 , wherein the recognizing the miniboard display is in real time with receiving of the live media stream.
11. The device of claim 6 , wherein the marking the positions of the static miniboard pixels is based on marking boundaries of recognized characters of static information portions.
12. The device of claim 6 , wherein the frames of the live media stream are latest received frames.
13. A method comprising:
determining, by a system including a processor, that a timed activity is visible in a live media stream;
receiving, by the system, a game clock value synchronized to the live media stream;
performing, by the system, optical character recognition on the live media stream according to the game clock value to identify game clock portions of frames of the live media stream:
determining, by the system, variances between sequences of the game clock portions of the frames of the live media stream by computing pixel intensity differences at pixel positions in the game clock portions;
identifying, by the system, static miniboard pixels and dynamic miniboard pixels according to the variances between the sequences of the game clock portions of the frames of the live media stream;
marking, by the system, positions of the static miniboard pixels and the dynamic miniboard pixels of the sequences of the game clock portions of the frames of the live media stream; and
recognizing, by the system, a miniboard display based on the marking of the positions.
14. The method of claim 13 , wherein the recognizing the miniboard display is in real time with receiving of the live media stream.
15. The method of claim 13 , wherein the frames of the live media stream are latest received frames.