Target Tracking Drift Detection Using Largest Response Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing target tracking methods suffer from tracker drift due to factors like blocking, motion blurring, and appearance changes, leading to inaccurate tracking and high computational costs, especially in real-time applications.
Innovation Solution
A target tracking method that uses a drift determining model based on the largest sample response values to quickly detect and correct tracker drift, improving accuracy and reducing computational burden by employing a probability density model for rapid response.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a detection module is added to the tracker to resolve tracker drift, then tracking accuracy is improved, but real-time application effect deteriorates due to excessively large computing amount
Solution Approach 1:
The patent extracts only the essential feature for drift detection - the largest sample response value - from the complex detection module. By monitoring changes in this single critical parameter rather than performing full detection algorithms, the system achieves drift detection with minimal computational overhead, resolving the contradiction between accuracy and real-time performance.
Solution Approach 2:
The patent performs preliminary action by pre-calculating and storing the largest sample response values during the tracking process. This allows the system to quickly compare current response values against historical data to detect drift, avoiding the need for computationally intensive detection algorithms during real-time operation.
2Reliability
If existing detection algorithms are used to detect tracker drift, then drift detection capability is improved, but detection speed deteriorates making it unsuitable for fast tracking
Solution Approach 1:
The patent extracts only the critical information needed for drift detection - the largest sample response value - from complex detection algorithms. By monitoring changes in this single parameter over time, the system achieves effective drift detection with extremely fast computation speed, suitable for real-time tracking applications.
Solution Approach 2:
The patent changes the detection parameter from complex image features or multiple algorithmic outputs to a single scalar value - the largest sample response value. This parameter transformation enables rapid comparison and drift detection while maintaining effectiveness, as drift manifests as significant changes in this response value.
Data Source
AI summary
In one embodiment, a target tracking method includes: receiving a current frame of picture including a target object; determining, based on a drift determining model, whether a tracker drifts for tracking of the target object in the current frame of picture, where the drift determining model is obtained through modeling based on largest values of responses values of a training sample used to train the drift determining model, where the training sample is collected from a training picture that includes the target object, where the response value of the sample is a value indicating a probability that the training sample is the target object in the training picture; and outputting a tracking drift result, where the tracking drift result includes: drift is generated for the tracking of the target object, or no drift is generated for the tracking of the target object.