Background Model for User Recognition in Depth Video
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately distinguishing a user's body from a complex background in depth video is challenging due to the complexity of separating the user from their surroundings, leading to erroneous associations or rejections in posture recognition systems.
Innovation Solution
A computer system with a hardware interface, computer-memory device, and update engine that maintains a background model using trusted coordinates from depth video, allowing the posture-recognition engine to ignore background portions and focus on the user's posture in front of the background, bounded by the trusted coordinates, which are updated based on observed pixel depth and brightness over specific durations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a background model is maintained using trusted coordinates from depth video, then the system can effectively separate the user from the background and reduce posture recognition errors, but the device complexity increases due to the need for maintaining and updating the background model
Solution Approach 1:
The system performs preliminary actions by maintaining a background model of the environment before posture recognition occurs. The background model includes trusted coordinates that are pre-established and stored in memory, allowing the system to quickly separate background from foreground (user) during posture recognition without performing complex real-time analysis of every pixel
Solution Approach 2:
The system segments the depth video into background portions and user portions by comparing observed pixel depths against trusted coordinates in the background model. This segmentation allows the posture recognition engine to focus only on regions containing the user, improving accuracy while reducing computational complexity by excluding background areas from detailed analysis
2Adaptability or versatility
If the background model is updated dynamically based on observed pixel depth, then the system adapts to environmental changes and maintains accuracy, but the loss of information increases due to potential errors in distinguishing real user movements from background changes
Solution Approach 1:
The background model is updated dynamically based on observed pixel depths from the depth camera. The update engine continuously adjusts trusted coordinates in the background model when it detects that observed pixels consistently fall behind or in front of existing trusted coordinates over specified durations, allowing the system to adapt to environmental changes such as objects being moved or added to the scene
Solution Approach 2:
The system uses feedback from the depth camera observations to continuously refine the background model. By monitoring whether observed pixel depths are in front of or behind trusted coordinates over time, the system receives feedback about environmental changes and adjusts the background model accordingly, improving adaptability while maintaining information accuracy through duration-based verification
3Reliability
If the system retains trusted coordinates when observed pixel depth is in front over a short duration but moves them when in front over a longer duration, then false updates are prevented, but the response time to detect real changes increases
Solution Approach 1:
The system employs periodic verification by checking observed pixel depths over specified durations before updating trusted coordinates. Short duration checks prevent false updates from transient movements, while longer duration checks confirm genuine environmental changes, creating a periodic verification mechanism that balances reliability and response time
Data Source
AI summary
A computer system comprises a hardware interface, a computer-memory device, an update engine, and a posture-recognition engine. The hardware interface is configured to receive depth video of an environment from a camera. The computer-memory device stores a background model of the environment preservable over a reboot of the computer system, the background model including a plurality of trusted coordinates derived from the depth video. The update engine is configured to update the background model, including moving a trusted coordinate to greater depth if an observed pixel is behind the trusted coordinate over a first duration, but retaining the trusted coordinate if the observed pixel depth is in front of the trusted coordinate over the first duration. The posture-recognition engine is configured to recognize posture of a user in front of a background portion of the video, which is bounded by the trusted coordinates of the background model.