Systems and methods for detection, prediction, and value estimation of activities and events
Systems and methods are described to analyze visual data and trajectory data from sports and games, to determine the content of the input data and detect actions and events, quantitatively measure their impact on the future activities and events and make predictions. In real-time, at every moment of the game, the probability of the next action/event in the game is estimated; the impact of the current state of the game on the future outcomes are predicted; and the current player action/match event is labelled. More specifically, given a predefined objective, a quantitative value is assigned to the input data based on the current observation characterizing the probability of reaching an objective, or incremental changes to that probability based on the observed state of the game from the video or trajectory data.
1 . A method for processing data to describe content of the data and to generate predictions associated with events occurring in the content, the method comprising:
receiving input data comprising at least one image from a video of a scene showing one or more entities at a corresponding time;
applying at least one descriptive and predictive model on the received data by way of a mapping function transforming the received input data into output data, the at least one descriptive and predictive model comprising the mapping function previously learnt from historical data; and
providing output data comprising an estimation of a likelihood of at least one event of interest occurring in the video or the scene, wherein the estimation is derived from visual content of the at least one image in the received input data.
2 . The method of claim 1 , further comprising:
using a training set comprising at least one predefined objective or a labeled event of interest and its corresponding time; and,
applying at least one machine learning or artificial intelligence technique to learn the mapping function from the training set, wherein the mapping function assigns a plurality of quantitative values to the input data to represent the content of the input data.
3 . The method of claim 1 , further comprising outputting at least one of: detected actions and events and their corresponding attributes at a current point in time, an impact of a current observation on reaching a predefined objective, a probability of occurrence of any action and/or event at a future time, an impact of the future action and/or event on reaching the predefined objective, and a probability of reaching the predefined objective in the future.
4 . The method of claim 1 , further comprising generating a probability distribution function characterizing future activities and events, given current observations.
5 . The method of claim 1 , further comprising generating a probability distribution function characterizing a predefined objective, given current observations.
6 . The method of claim 1 , further comprising generating the location data of the entities from current visual observations.
7 . The method of claim 1 , further comprising predicting future locations of the entities.
8 . The method of the claim 1 , further comprising generating the output data in real-time for sports betting and/or media applications.
9 . The method of claim 1 , further comprising generating at least one quantitative metric to rank players or teams and assess player performance or team performance in games, measure an effect of a player or team activity in the game, a player contribution to a future game, or a player or team efficiency in achieving a particular outcome.
10 . The method of claim 1 , further comprising generating at least one quantitative metric to predict player performance and game outcomes.
11 . The method of claim 1 , further comprising outputting at least one label to categorize an individual or a group activity in the visual data.
12 . The method of claim 1 , further comprising generating at least one quantitative metric to predict an outcome of a sport match.
13 . The method of claim 1 , further comprising using artificial neural networks to represent the mapping function.
14 . The method of claim 1 further comprising using reinforcement learning techniques to learn the mapping functions.
15 . A non-transitory computer readable medium storing computer executable instructions for processing data to describe content of the data and to generate predictions associated with events occurring in the content, comprising instructions for:
receiving input data comprising at least one image from a video of a scene showing one or more entities at a corresponding time;
applying at least one descriptive and predictive model on the received data by way of a mapping function transforming the received input data into output data, the at least one descriptive and predictive model comprising the mapping function previously learnt from historical data; and
providing output data comprising an estimation of a likelihood of at least one event of interest occurring in the video or the scene, wherein the estimation is derived from visual content of the at least one image in the received input data.
16 . The computer readable medium of claim 15 , further comprising instructions for:
using a training set comprising at least one predefined objective or a labeled event of interest and its corresponding time; and,
applying at least one machine learning or artificial intelligence technique to learn the mapping function from the training set, wherein the mapping function assigns a plurality of quantitative values to the input data to represent the content of the input data.
17 . The computer readable medium of claim 15 , further comprising instructions for outputting at least one of: detected actions and events and their corresponding attributes at a current point in time, an impact of a current observation on reaching a predefined objective, a probability of occurrence of any action and/or event at a future time, an impact of the future action and/or event on reaching the predefined objective, and a probability of reaching the predefined objective in the future.
18 . The computer readable medium of claim 15 , further comprising instructions for generating a probability distribution function characterizing future activities and events, given current observations.
19 . The computer readable medium of claim 15 , further comprising instructions for generating a probability distribution function characterizing a predefined objective, given current observations.
20 . A device comprising a process and memory, the memory storing non-transitory computer readable instructions for processing data to describe content of the data and to generate predictions associated with events occurring in the content, comprising instructions for:
receiving input data comprising at least one image from a video of a scene showing one or more entities at a corresponding time;
applying at least one descriptive and predictive model on the received data by way of a mapping function transforming the received input data into output data, the at least one descriptive and predictive model comprising the mapping function previously learnt from historical data; and
providing output data comprising an estimation of a likelihood of at least one event of interest occurring in the video or the scene, wherein the estimation is derived from visual content of the at least one image in the received input data.