Automatic Image Segmentation Using GMM-WD Tri-Map Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic segmentation of static images into foreground, background, and unknown regions is challenging due to under-constrained natural image matting, requiring user interaction like scribbles or tri-maps for accurate labeling.
Innovation Solution
The method employs Gaussian Mixture Models with weighted data (GMM-WD) to identify and segment foreground, background, and unknown regions in images, allowing for automatic generation of a tri-map without user input by applying GMM-WD iteratively to refine body and background regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If single-image matting approach is used with linear combination of foreground and background colors, then the method can handle natural images without background restrictions, but the problem becomes inherently under-constrained requiring user interaction
Solution Approach 1:
The patent segments the image into three distinct regions: foreground, background, and unknown regions. This segmentation is achieved by applying GMM-WD to identify foreground pixels, then identifying background pixels from the remaining areas, and finally determining unknown regions by exclusion. This three-way segmentation resolves the under-constrained nature of single-image matting by providing explicit regional information without requiring user interaction.
Solution Approach 2:
The patent implements automatic tri-map generation through self-service mechanisms. The GMM-WD algorithm automatically identifies foreground and background regions, and the unknown regions are automatically determined by excluding the identified foreground and background from the entire image. This self-service approach eliminates the need for user-provided tri-maps or scribbles while maintaining adaptability to natural images.
2Measurement precision
If user interaction such as scribble interface or tri-map is required for complete labeling, then segmentation accuracy can be achieved, but automation cannot be attained
Solution Approach 1:
The patent achieves automatic segmentation through self-service mechanisms where the GMM-WD algorithm independently identifies foreground and background regions without user input. The system automatically generates the tri-map by excluding identified foreground and background pixels from the entire image, thereby attaining complete automation while maintaining segmentation accuracy through the mathematical rigor of GMM-WD.
Solution Approach 2:
The patent replaces the mechanical interaction system (user scribbles or tri-map drawing) with an automated computational system based on GMM-WD. Instead of requiring manual pixel labeling, the system uses Gaussian Mixture Models with weighted data to automatically classify pixels into foreground, background, and unknown regions, substituting human mechanical action with algorithmic processing.
3Extent of automation
If GMM-WD is applied iteratively to refine foreground and background regions, then automatic tri-map generation is achieved, but computational complexity increases
Solution Approach 1:
The patent divides the computational process into distinct sequential stages: first applying GMM-WD to identify foreground regions, then applying GMM-WD again to identify background regions from the remaining areas, and finally determining unknown regions by exclusion. This segmentation of the computational process manages complexity by breaking down the automatic tri-map generation into manageable, systematic steps rather than requiring a single complex algorithm.
Data Source
AI summary
A method for achieving segmentation of a picture according to one aspect of the present invention comprises: determining a first foreground of a picture based on a predetermined mask; applying Gaussian Mixture Models with weighted data (GMM-WD) to the first foreground to generate a second foreground; determining a first background of the picture based on the second foreground; applying the GMM-WD to the first background to generate a second background; and determining an unknown region based on the second background and the second foreground.