Pattern Matching System with Time-Variant Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pattern matching systems face challenges in accurately and safely matching patterns that include features changing over time, such as fingerprints with wrinkles or scars, especially when the time interval between pattern registration and comparison is long, leading to potential misidentification or erroneous template updates.
Innovation Solution
A pattern matching system that separates features into those not changing with time and those changing with time, performing distinct matching processes for each to ensure accurate and safe comparison, including a template storage unit, feature extraction unit, and matching unit that handles both types of features separately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If features changing with time are used for matching, then matching accuracy is improved, but reliability deteriorates when time interval is long
Solution Approach 1:
The patent segments features into two categories: features that change with time (first features) and features that do not change with time (second features). By separating these feature types, the system can apply different matching strategies to each category, thereby achieving both high matching accuracy using time-sensitive features and reliable matching using time-invariant features when time intervals are long.
2Reliability
If only features not changing with time are used for matching, then reliability is improved, but matching accuracy deteriorates
Solution Approach 1:
The patent merges the matching results of two different feature types: time-invariant features (second features) and time-variant features (first features). By combining these two matching processes, the system achieves both reliability from time-invariant features and high accuracy from time-variant features, resolving the contradiction between the two.
3Reliability
If allowable error is set large to accommodate time changes, then reliability is improved, but matching accuracy deteriorates
Solution Approach 1:
The patent applies different quality standards (allowable error thresholds) to different feature types. For time-invariant features, a strict matching criterion is applied, while for time-variant features, a more lenient criterion is used. This local differentiation of matching strictness allows the system to maintain high overall accuracy while accommodating time changes in specific feature types.
Data Source
AI summary
A problem to be solved is to provide a pattern matching system and the like which can match patterns including time change with high accuracy and safety. The problem can be solved by a pattern matching system which includes a template storage unit, a feature extraction unit, and a matching unit. The template pattern stores a template pattern. The feature extraction unit extracts features of an input pattern. The matching unit performs a first matching in which a first feature not changing with time among the features of the input pattern is matched with the template pattern and performs a second matching other than the first matching, in which a second feature changing with time among the features of the input pattern is matched with the template pattern.