Sleeping Object Detection in Video Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video analytics systems face challenges in detecting and tracking sleeping objects, as background subtraction methods fail to maintain object detection when objects become stationary, causing them to fade into the background and lose tracking.
Innovation Solution
The system employs a method to identify and maintain blob trackers by determining the type of tracker associated with objects that have transitioned into the background model, modifying tracker characteristics based on interactions with other trackers, and updating appearance models to ensure continuous tracking of sleeping objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If background subtraction is used to detect motion objects, then moving objects can be detected and tracked, but sleeping objects (stationary objects) cannot be detected as they fade into the background
Solution Approach 1:
The system dynamically adjusts tracker characteristics based on object state. When an object is detected as sleeping (stationary), the system transitions it to a special sleeping tracker state that prevents background model updates, allowing continuous tracking despite the object being stationary. This dynamic state adjustment resolves the contradiction by adapting the tracking mechanism to the object's motion state.
Solution Approach 2:
The system changes key parameters of the background subtraction model when sleeping objects are detected. Specifically, it modifies the background model update rule to exclude pixels from sleeping objects, preventing them from being absorbed into the background. This parameter change allows the system to maintain detection reliability for both moving and stationary objects.
2Speed
If background subtraction adapts quickly to local changes, then moving objects are tracked accurately, but sleeping objects transition to background model and are lost
Solution Approach 1:
The system applies different quality rules to different regions of the image based on object state. For sleeping objects, it creates a local exclusion zone around the object where background model updates are prevented. This local quality adjustment allows fast adaptation in most regions while maintaining tracking reliability for stationary objects by protecting their regions from background model contamination.
3Productivity
If standard blob tracking is used, then moving objects are tracked, but the system complexity increases when handling interactions between sleeping objects and other objects
Solution Approach 1:
The system performs preliminary classification of trackers into standard and sleeping types. By pre-identifying sleeping objects and assigning them special tracker characteristics before interactions occur, the system simplifies subsequent interaction handling. This preliminary action prevents the need for complex real-time analysis of tracker-object interactions, maintaining productivity while reducing operational complexity.
Data Source
AI summary
Methods, apparatuses, and computer-readable media are provided for maintaining blob trackers for video frames. For example, a first blob tracker maintained for a current video frame is identified. The first blob tracker is associated with a blob detected in one or more video frames. The blob includes pixels of at least a portion of a foreground object in the one or more video frames. It is determined that the first blob tracker is a first type of tracker. Trackers having the first type are associated with objects that have transitioned at least partially into a background model (referred to as sleeping objects and sleeping trackers). One or more interactions are identified between the first blob tracker and at least one other blob tracker. The at least one other blob tracker can be the first type of tracker or can be a second type of tracker that is not a sleeping tracker (the second type of tracker is not associated with an object that has transitioned at least partially into the background model. A characteristic of the first blob tracker can then be modified based on the identified one or more interactions. Modifying the characteristic of the first blob tracker can include transitioning the first blob tracker from the first type of tracker to the second type of tracker, updating an appearance model of the first blob tracker, and/or other suitable characteristic of the first blob tracker.