Occluded Object Tracking via Connected Group Motion Tracks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking systems struggle with occlusions when objects move in groups, leading to the loss of object tracks and the inability to re-identify temporarily occluded objects, especially in multi-camera settings.
Innovation Solution
A method where inactive object tracks of temporarily occluded objects are connected to active object tracks of associated objects moving together, using a less strict similarity requirement when the detected object is spatially proximate, allowing for resumed tracking and re-activation of the occluded object's track.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a strict similarity requirement is used for object re-identification, then the accuracy of track association is improved, but the likelihood of losing tracks during occlusion increases
Solution Approach 1:
The patent applies dynamics by making the similarity requirement adaptive rather than static. The system dynamically adjusts the similarity threshold based on whether an object is currently occluded or visible. When objects are visible, a strict similarity requirement ensures accurate associations. When occlusion is detected, the system relaxes the similarity threshold to maintain track continuity, allowing potential matches that would otherwise be rejected. This dynamic adjustment resolves the contradiction between accuracy and continuity.
Solution Approach 2:
The patent changes the parameter of similarity threshold based on the occlusion state of objects. By monitoring whether objects are visible or occluded in consecutive frames, the system adjusts the similarity requirement parameter accordingly. This parameter change allows the system to maintain high association accuracy when possible while ensuring track survival during occlusion events, effectively resolving the contradiction between measurement precision and reliability.
2Device complexity
If objects move in groups, then the complexity of tracking is reduced, but the occurrence of occlusions between objects increases
Solution Approach 1:
The patent applies preliminary action by proactively identifying groups of objects that move together and pre-establishing their spatial relationships. When objects are detected moving in proximity with consistent relative positions, the system identifies them as a group and prepares for potential occlusions. This preliminary identification allows the system to anticipate occlusion events and adjust tracking parameters in advance, maintaining group tracking simplicity while managing occlusion risks.
Solution Approach 2:
The patent uses visible objects in a group as intermediaries to maintain tracks of occluded objects. When one object occludes another, the visible object serves as a mediator that provides spatial context and motion information. The system uses the visible object's trajectory and position to infer the likely position of the occluded object, allowing continuous tracking without direct observation. This intermediary approach reduces the harmful effects of occlusion while maintaining the simplicity of group tracking.
3Reliability
If an object track is kept inactive for a long time, then the possibility of re-identification increases, but computational resources are wasted
Solution Approach 1:
The patent applies preliminary action by proactively identifying occlusion events and immediately transitioning tracks to an inactive state with relaxed similarity requirements. Rather than waiting for the timeout period to expire, the system detects occlusion through analysis of object visibility in consecutive frames and takes preliminary action to preserve the track. This allows the system to maintain fewer active tracks during occlusion events, reducing computational overhead while ensuring re-identification capability when objects reappear.
Data Source
AI summary
An apparatus and a method for tracking a temporarily occluded object in image frames of objects and associating detected instances into tracks. An active first track of a first object and an active second track of a second object are identified, which objects are moving together. Upon determination that the first track, but not the second track, has become inactive connecting it to the second track. Thereafter, a detected object instance is associated with an inactive track when a similarity score between the detected instance and the inactive track fulfills a similarity requirement. The requirement is less strict when the instance is spatially proximate to the second track and when the detected instance is compared to the inactive first track connected to the second track as compared to when the detected instance is compared to an unconnected inactive track.