Motion Detection Using Grey Relational Analysis for VBR Video
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional background subtraction methods for motion detection in video streams struggle with accurately identifying moving objects in variable-bit-rate (VBR) video streams due to bit rate changes, leading to misinterpretation of motion signals as background signals or vice versa, especially in real-world networks prone to congestion and bandwidth constraints.
Innovation Solution
A motion detection method based on grey relational analysis that establishes a multi-quality background model by calculating Euclidean distances and grey relational coefficients to differentiate between bit rate changes, employs block-based and pixel-based detection procedures to generate a binary motion mask, and uses entropy calculation to update the background model in response to luminance changes, while providing an interface for setting detection sensitivity to minimize false detections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional background subtraction methods are used in VBR video streams, then implementation is simple and calculated amount is moderate, but detection accuracy deteriorates due to bit rate changes causing misinterpretation of motion signals
Solution Approach 1:
The patent applies dynamics by making the background model adaptive to changing bit rate conditions. The system dynamically adjusts the background model based on detected bit rate changes, allowing it to adapt its behavior to match the current video quality conditions rather than using a fixed model. This resolves the contradiction by maintaining simple implementation while improving detection accuracy through dynamic adaptation.
Solution Approach 2:
The patent changes the parameter of background model characteristics based on bit rate conditions. When bit rate changes are detected, the system modifies the background model parameters to account for the new video quality conditions. This allows the same simple background subtraction algorithm to work accurately across varying bit rates by adjusting its parameters rather than changing the fundamental approach.
2Measurement precision
If background model is updated frequently to adapt to bit rate changes, then detection accuracy improves, but system stability deteriorates due to false detections from rapid updates
Solution Approach 1:
The patent applies preliminary action by detecting bit rate changes before they cause detection errors. The system proactively identifies when bit rate conditions are changing and pre-adjusts the background model accordingly, preventing false detections rather than reacting to them after they occur. This timing strategy improves accuracy without causing instability from reactive updates.
Solution Approach 2:
The patent uses feedback by continuously monitoring video stream conditions and using this information to control background model updates. The system only updates the background model when specific conditions are met (detected bit rate changes), creating a feedback-controlled update mechanism that balances accuracy improvement with system stability by avoiding unnecessary updates.
3Object-affected harmful factors
If background model is updated to match current background signal, then false detections from background changes are reduced, but missed detections increase when motion signals are present during updates
Solution Approach 1:
The patent introduces bit rate change detection as an intermediary mechanism between the video stream and the background model update process. This intermediary layer analyzes the video conditions and only permits background model updates when appropriate (during actual bit rate changes), preventing updates that would cause missed detections while still allowing updates that reduce false detections from background changes.
Data Source
AI summary
A motion detection method determines bit-rate changes of input pixels of a video frame by a grey relational analysis technique to establish a multi-quality background model, detects moving objects by two-stage block-based and pixel-based detection procedures to generate a binary motion mask, detects luminance changes of the video frame by entropy calculation to timely update the background model, provides a setting interface for a user to set a detection sensitivity, and examines false detections of the binary motion mask. Therefore, it can correctly interpret moving objects in VBR video streams, implement more accurate and complete motion detection, eliminate the influence of luminance changes, increase the detection accuracy, and decrease false detections.