Object Tracking Apparatus Dynamic Removal Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in simultaneously removing erroneously registered tracking targets and continuously tracking objects that are temporarily undetected in moving images, as they struggle to balance early removal of unnecessary targets with maintaining tracking of objects hidden behind other objects.
Innovation Solution
An information processing apparatus that includes detection means for identifying tracking candidates, corresponding degree calculation means to assess the probability of candidate objects matching registered targets, and decision means to determine whether to continue or delete tracking based on calculated evaluation values, allowing for early removal of unnecessary targets and continuous tracking of temporarily undetected objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the computer performs processing to remove an undetected moving object from the tracking target, then erroneous registrations are eliminated, but objects that are temporarily hidden behind other objects are incorrectly removed
Solution Approach 1:
The patent applies dynamics by making the tracking target management adaptive and changeable over time. The computer dynamically adjusts the tracking state based on detection results, allowing the tracking target list to be flexibly updated. When an object is temporarily undetected (hidden), the system maintains its tracking target status, but when an object is consistently undetected (erroneous registration), the system removes it. This dynamic adaptation resolves the contradiction between removing erroneous targets and maintaining hidden targets.
Solution Approach 2:
The patent changes the parameter of tracking target management from a static removal criterion to a dynamic state-based criterion. Instead of simply removing all undetected objects, the system introduces a detection state parameter that distinguishes between temporary occlusion and erroneous registration. By monitoring detection continuity and object behavior parameters, the system accurately determines whether to maintain or remove tracking targets, resolving the contradiction between detection accuracy and tracking reliability.
2Duration of action of stationary object
If the computer waits for continuous non-detection before removing a tracking target, then temporarily hidden objects are maintained, but erroneous registrations persist longer
Solution Approach 1:
The patent applies preliminary action by proactively monitoring detection patterns and object behavior before final removal decisions are made. The system preliminarily identifies potentially erroneous registrations by analyzing detection continuity, object movement patterns, and tracking history. This preliminary analysis allows the system to distinguish between temporarily hidden objects and erroneous registrations earlier in the process, enabling timely and accurate removal decisions that maintain reliability while appropriate duration is granted to legitimate hidden objects.
Data Source
AI summary
Provided is a technology which enables early removal from a tracking target of an object unnecessary to be tracked and which, together therewith, when the tracking target is temporarily in a state of being not detected, enables continuous tracking of the relevant tracking target without removing it from the tracking target.A detection unit 2 detects, as a tracking candidate, an object included in an image by using image processing. A corresponding degree calculation unit 3 calculates a corresponding degree which is index representing a probability that the detected tracking candidate corresponds to the tracking target registered as an object to be tracked. When having determined that it is unnecessary to track the tracking target based on an evaluation value calculated by using the corresponding degree, a decision unit 4 deletes the registration of the object to be tracked with respect to the relevant tracking target.