Video Object Extraction Using Multiple Background Templates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting objects of interest from video images during sport events using background subtraction techniques are inconsistent and dependent on various parameters, leading to variable results due to differences in background nature and object movement, which affects the quality of object extraction.
Innovation Solution
A method and apparatus that utilize multiple background templates derived using different background subtraction techniques and parameters, calculating differences between video images and these templates, and applying rules to dynamically combine and balance these differences for optimized object extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single background template is used for object extraction, then the processing is simple and fast, but the extraction quality is inconsistent and variable
Solution Approach 1:
The patent combines multiple background templates (first background template from static analysis and second background template from motion analysis) into a unified object extraction process. By merging the results from different template analyses, the system achieves more consistent and reliable object extraction compared to using a single template, resolving the contradiction between extraction quality and processing complexity.
Solution Approach 2:
The patent changes the parameter of background template representation by using multiple templates with different characteristics (static vs. motion-based) instead of a single template. This parameter change allows the system to adapt to varying background conditions and improve extraction quality while managing complexity through structured processing.
2Reliability
If multiple background templates are used for object extraction, then the extraction quality and consistency are improved, but the processing complexity increases
Solution Approach 1:
The patent segments the background modeling process into distinct components: a first background template derived from static image analysis and a second background template derived from motion analysis. This segmentation allows each template to specialize in capturing different aspects of the background, improving reliability while organizing the complexity into manageable, separate processing streams.
Solution Approach 2:
The patent introduces an intermediary processing stage that combines results from multiple background template analyses. This intermediary step integrates the strengths of different template approaches and applies logical rules to synthesize the final object extraction result, thereby improving consistency while systematically managing the increased processing complexity.
3Speed
If background subtraction is applied to detect moving objects, then the detection speed is fast, but the results are affected by background variations such as weather changes
Solution Approach 1:
The patent implements a dynamic background modeling approach that adapts to changing conditions. The system maintains multiple background templates and selectively applies them based on current scene characteristics, allowing the detection process to adapt to background variations like weather changes while preserving fast detection speeds through efficient template selection and combination.
Solution Approach 2:
The patent changes the parameter of background representation by using multiple templates with different temporal and spatial characteristics. This allows the system to adapt to varying background conditions (such as weather changes) by selecting or combining appropriate templates, thereby improving versatility without significantly compromising detection speed.
Data Source
AI summary
A computer implemented method of object extraction from video images, the method comprising steps a computer is programmed to perform, the steps comprising: receiving a plurality of video images, deriving a plurality of background templates from at least one of the received video images, calculating a plurality of differences from an individual one of the received video images, each one of the differences being calculated between the individual video image and a respective and different one of the background templates, and extracting an object of interest from the individual video image, using a rule applied on the calculated differences.