Spatiotemporal Motion Pattern Analysis Using Partial Tracking Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI systems struggle to analyze motion patterns in low-scoring, continuous sports like soccer and hockey, as they rely on segmented plays, which are not easily identifiable, and often lack complete player tracking data, making it difficult to build libraries of successful and unsuccessful plans.
Innovation Solution
A method that characterizes team behaviors using partial tracking data, such as ball-action data, by partitioning input data into spatiotemporal segments, generating representations based on criteria, and computing metrics, allowing for the creation of entropy maps to measure predictability, without relying on pre-defined libraries or full player tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the system segments game data into discrete plays (as in tennis), then it can build a library of successful and unsuccessful plans, but this approach fails for continuous, low-scoring sports like soccer and hockey where such segmentation is not easily identifiable
Solution Approach 1:
The patent segments continuous sports data into discrete spatiotemporal plays by identifying key events (goals, shots, tackles) and extracting meaningful segments between these events. This allows the system to apply play-library methodologies to continuous sports by creating artificial but meaningful segmentation based on game-critical moments rather than relying on pre-existing discrete segments.
Solution Approach 2:
The system changes the parameters used to define play segments by incorporating multiple criteria such as time thresholds, spatial boundaries, and event-based markers. This allows flexible adaptation of the segmentation approach to different sport types and scoring patterns, making the system versatile across both segmented and continuous sports.
2Measurement precision
If the system uses complete player tracking data, then it can accurately analyze team behaviors and motion patterns, but tracking data is costly and difficult to obtain, making only partial data (such as ball tracking) available
Solution Approach 1:
The patent extracts and utilizes only the essential tracking data that is available (such as ball position and key player events) rather than requiring complete player tracking data. By focusing on extracting meaningful information from partial data sources and combining them with event-based annotations, the system achieves accurate team behavior analysis without the need for costly complete tracking infrastructure.
Solution Approach 2:
The system introduces event-based annotations and spatiotemporal segmentation as intermediaries that bridge the gap between partial tracking data and comprehensive team behavior analysis. These intermediaries allow the system to infer player movements and team patterns even when direct player tracking data is unavailable, using ball tracking and event markers as mediators.
3Adaptability or versatility
If the system analyzes continuous sports without pre-defined plan libraries, then it can be applied to more sport types, but it lacks the means or labels to form such libraries automatically
Solution Approach 1:
The system performs preliminary automated segmentation and labeling of plays based on detected key events and spatiotemporal patterns. By automatically creating initial play segments and assigning labels based on event types and outcomes, the system prepares the data structure needed for plan library formation without requiring manual intervention, thus reducing construction complexity while maintaining adaptability.
Solution Approach 2:
The system enables automatic plan library construction by having the analysis system generate its own training data through automated play segmentation and labeling. The system uses its own event detection and pattern recognition capabilities to create labeled plays that can serve as the foundation for plan libraries, making the library construction process self-service rather than requiring external manual labeling.
Data Source
AI summary
Techniques are described to characterize motion patterns of a group of agents engaging in an activity. An analysis system receives input data associated with spatial and temporal information of at least one element of interest associated with the activity, where the object of interest may be a ball, person, animal or any other object in motion. The analysis system partitions the input data into a plurality of spatiotemporal segments and generates one or more representations of one or more sets of segments of the plurality of spatiotemporal segments based on one or more criteria. The analysis system computes a metric, such as an entropy value, for each of the one or more representations. Partial tracing data, such as ball movements in a sporting event, may be created using an inexpensive input device, such as a tablet computer, making the disclosed techniques available for a wide range of events and activities.