Template Matching Using Normalized Feature Membership Indicators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital image processing techniques for template matching lack a comprehensive method to preserve contextual information and accurately compare entities across different alignment shifts using various template matchers, leading to inconsistencies in feature measurement and alignment evaluation.
Innovation Solution
The system generates matcher-specific feature values and feature membership indicators by executing processor instructions that use multiple template matchers with different mathematical comparison techniques, distance metrics, and feature categories, normalizing and weighting these values to produce a normalized accumulated feature membership indicator that represents the best match alignment shift, thereby preserving contextual information and ensuring relative alignment consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple template matchers with different mathematical comparison techniques are used, then measurement precision and reliability are improved, but device complexity increases
Solution Approach 1:
The patent divides the template matching process into multiple independent matchers, each handling specific feature categories (edges, corners, BLOBs, ridges) with appropriate mathematical comparison techniques. This segmentation allows each matcher to specialize in detecting particular image features, improving overall measurement precision while maintaining manageable complexity through modular design.
Solution Approach 2:
The patent combines results from multiple template matchers by generating feature membership indicators for each matcher and accumulating them into a normalized accumulated feature membership indicator. This merging process integrates diverse measurement perspectives (different matchers) to produce a more reliable and precise final alignment measurement, resolving the contradiction between using multiple matchers and system complexity.
2Reliability
If feature membership indicators are generated by normalizing matcher-specific feature values, then reliability is improved through managing modeling uncertainties, but loss of information may occur during normalization
Solution Approach 1:
The patent transforms matcher-specific feature values into feature membership indicators through normalization, changing the parameter scale to a unified range (typically 0 to 1). This parameter transformation enables reliable comparison and accumulation of features from different matchers with different mathematical comparison techniques, improving reliability by managing modeling uncertainties while preserving essential relative information through the normalization process.
Data Source
AI summary
Systems, articles of manufacture, and methods for template matching at least one template image and an image to be searched using a plurality of template matchers at at least one alignment shift.