Dynamic Feature Combination for Geometric Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current geometric pattern matching algorithms face difficulties in accurately locating objects of interest in images, particularly due to issues with feature characterization and matching efficiency.
Innovation Solution
The method involves a learning phase where geometric features of a template image are analyzed to create a model with a relationship tree, and a matching phase where these features are dynamically combined to locate instances of the template in a target image, using a hierarchical approach to extract and match features, and update matches based on strength and spatial relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional pattern matching algorithms are used to locate template images, then the search process can be completed, but the accuracy and efficiency of detecting objects of interest deteriorates due to insufficient feature characterization
Solution Approach 1:
The patent segments the template image into multiple distinct geometric features (lines, circles, arcs, corners) and characterizes each feature separately with specific properties. This segmentation allows the system to focus on individual feature types and their relationships, improving detection accuracy while managing complexity through organized feature classification.
Solution Approach 2:
The patent changes the parameters used for feature characterization by extracting specific geometric properties (coordinates, radii, angles, lengths) from the template image. These parameter changes enable more precise feature representation and matching, transforming the approach from general pixel-based comparison to specific geometric parameter matching.
2Measurement precision
If comprehensive feature extraction is performed on the template image, then the matching accuracy improves, but the processing time and computational load increases
Solution Approach 1:
The patent performs preliminary action by extracting and characterizing all geometric features from the template image before the actual matching process. During this learning phase, the system pre-computes feature relationships, spatial configurations, and geometric properties, storing them for efficient retrieval during matching. This preliminary characterization reduces processing time during the actual search phase.
Solution Approach 2:
The patent segments the feature extraction process into distinct geometric feature types (lines, circles, arcs, corners) and processes each type with specialized algorithms. This segmentation allows for more efficient processing of each feature category separately, reducing overall computational load while maintaining comprehensive feature characterization for accurate matching.
3Reliability
If multiple feature types are used in pattern matching, then the robustness of the matching algorithm improves, but the device complexity and algorithm sophistication increases
Solution Approach 1:
The patent segments the feature set into distinct geometric categories (lines, circles, arcs, corners) with specific properties and relationships for each type. This segmentation provides a structured framework that improves robustness by ensuring comprehensive coverage of different geometric features while managing algorithm complexity through organized classification and type-specific processing rules.
Solution Approach 2:
The patent creates a universal feature characterization framework that handles multiple geometric feature types through a common structure. The system uses a unified approach to extract, represent, and match features of different types, allowing the algorithm to be robust across various geometric configurations while avoiding the complexity of entirely separate processing paths for each feature type.
4Measurement precision
If the template image is characterized with detailed geometric features, then the ability to locate precise instances improves, but the amount of data to be processed increases
Solution Approach 1:
The patent extracts only the essential geometric features and their critical properties from the template image, removing redundant information. By taking out only the necessary geometric parameters (coordinates, dimensions, angles, relationships) rather than processing the entire image data, the system achieves precise location capability while minimizing the volume of data that needs to be stored and processed during matching.
Data Source
AI summary
A geometric pattern matching method to locate instances of a template image in a target image using a plurality of models with different feature combinations. A learning phase involves learning geometric information about the template image and creating a model for each of a plurality of different respective parent features. Each model comprises a relationship tree from the respective parent feature to one or more other child features, and hence each model may have different feature combinations. In a matching phase, the method may examine a plurality of different models to determine matches in the target image being analyzed. The matching phase may select different models dynamically based on prior matching results.