Multi-User Touch Trajectory Matching With Deferred Acceptance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing touch screen technologies struggle to accurately distinguish between multiple users' touch points, leading to inaccurate touch trajectory generation and user operation misjudgment, especially when multiple individuals interact with the screen simultaneously.
Innovation Solution
A method utilizing a deferred-acceptance algorithm to match touch points by correlating actual touch points with predicted points, enhancing accuracy through a deferred-acceptance algorithm, such as the Gale-Shapley algorithm, to stabilize touch point matching and generate precise touch trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional touch point association method is used, then the process is simple, but the accuracy of touch trajectory generation deteriorates when multiple users interact simultaneously
Solution Approach 1:
The patent applies preliminary action by predicting the next touch point position based on the current touch trajectory before actual touch detection occurs. This prediction is made in advance for multiple users simultaneously, creating a reference framework that guides subsequent touch point association, thereby improving accuracy without requiring complex real-time differentiation processes
Solution Approach 2:
The patent implements feedback through an iterative association process where touch points from current frames are matched with predicted touch points from previous frames. The system continuously refines the matching based on spatial proximity and temporal continuity, adjusting associations until optimal pairing is achieved, thus enhancing measurement precision through iterative refinement
2Reliability
If closest touch point association is used, then the computation is fast, but user operation identification becomes inaccurate when multiple users touch simultaneously
Solution Approach 1:
The patent applies segmentation by separating touch points into distinct sequences for different users based on spatial proximity and temporal continuity. Instead of treating all touch points uniformly, the system divides them into user-specific trajectories by matching predicted positions with actual detections, allowing accurate identification even when multiple users interact simultaneously
Solution Approach 2:
The patent implements dynamics by making the touch point association process adaptive rather than static. The system dynamically adjusts the association criteria based on real-time touch positions, velocities, and predicted trajectories, allowing flexible differentiation between users whose touch paths may overlap or intersect, thus maintaining high reliability without fixed computational constraints
Data Source
AI summary
Provided a method for matching touch points, a method for obtaining a touch trajectory, and products. The method for matching touch points includes: obtaining at least one touch point sequence, where the at least one touch point sequence including a first touch point sequence; obtaining at least one predicted point sequence, where the at least one predicted point sequence includes a first predicted point sequence which includes Q predicted points associated with a first actual touch point, the first actual touch point being one of the actual touch points, and at least one of the actual touch points being in a one-to-one correspondence with the at least one predicted point sequence; and matching the actual touch points in the at least one touch point sequence with the predicted points in the at least one predicted point sequence by a deferred-acceptance algorithm to obtain a matching result.


