Object Tracking Control Using Attribute-Based Occlusion Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking techniques in network cameras face challenges in maintaining detection accuracy when a target object becomes invisible due to being hidden behind another object, leading to visibility degradation and ghosting issues, especially when regions of interest (ROIs) with different code amounts overlap.
Innovation Solution
An image processing system that includes a tracking unit, attribute determination unit, and tracking control unit to detect and track objects, determine the attributes of overlapping objects, and control the tracking process based on the attributes and positional changes, allowing for continued or terminated tracking depending on the likelihood of the object reappearing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pattern matching method is used to detect objects in video, then detection can be performed using human body features, but detection accuracy decreases when target object is hidden behind other objects
Solution Approach 1:
The system maintains tracking information continuously even when the target object is occluded by other objects. The tracking unit preserves detection results from previous frames and continues tracking without interruption, ensuring uninterrupted monitoring of the target object's position and movement throughout the video sequence.
Solution Approach 2:
The system performs preliminary detection and tracking setup before occlusion occurs. By establishing tracking information in advance when the target is visible, the system prepares detection data that can be maintained during subsequent occlusion periods, enabling faster recovery when the target becomes visible again.
2Loss of substance
If dynamic ROI is applied to reduce bit rate, then data amount is reduced, but visibility of background objects degrades when ROIs with different code amounts overlap
Solution Approach 1:
The system applies different code amounts to different spatial regions based on object importance. Foreground objects receive higher code amounts for better quality, while background objects receive lower code amounts. This local differentiation optimizes bit rate distribution while maintaining visibility of important objects.
Solution Approach 2:
The system determines the front-and-back relationship between objects in advance before encoding. By identifying which objects are in front and which are in the background beforehand, the system can appropriately allocate code amounts to maintain visibility of background objects that may be partially occluded, preventing complete loss of important visual information.
3Reliability
If tracking is continued when object is hidden behind another object, then tracking information is maintained, but tracking resources are wasted when object will not reappear
Solution Approach 1:
The system uses the front-and-back relationship information as feedback to dynamically adjust tracking behavior. When a background object is occluded by a foreground object, the system feedback-indicates that tracking should be maintained. When no foreground object occludes the target, the system feedback-indicates that tracking should be terminated, optimizing resource usage based on actual visibility conditions.
Solution Approach 2:
The tracking behavior is made dynamic rather than static. The system continuously evaluates the front-and-back relationship between objects and adjusts tracking continuation or termination accordingly. This dynamic adaptation allows the system to maintain tracking only when necessary, balancing information continuity with resource conservation.
Data Source
AI summary
An image processing apparatus detects a tracking target object in an image, executes tracking processing to track the object, determines whether an attribute of an object detected from the image is a predetermined attribute, identifies, when a first state in which the object is detected changes to a second state with the object not detected, a given object included in the image and positioned at least partially in front of the object in the second state, based on a position of the object in the first state, controls the tracking processing, based on a determination whether an attribute of the given object is the predetermined attribute, and determines, based on the result, whether to continue the tracking processing. The tracking processing continues until at least a predetermined time has elapsed when a determination to continue the tracking processing on the tracking target object for the predetermined time has been made.