Transform Pyramiding for Fingerprint Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital fingerprinting systems struggle with accurately authenticating objects due to changes in viewing angle, position, illumination, and shape changes between successive digital fingerprint acquisitions, leading to failed authentication even when the object remains the same.
Innovation Solution
The implementation of a transform pyramiding process that approximates the distortion between digital fingerprints using similarity, affine, and homographic transforms, allowing for efficient matching and authentication by characterizing the overall transform as a series of local region transforms linked together.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional digital fingerprinting systems are used to authenticate objects, then the authentication process is simple, but the accuracy fails when objects undergo changes in viewing angle, position, illumination, or shape
Solution Approach 1:
The patent divides the digital fingerprint matching process into multiple levels of transform complexity (similarity, affine, homographic transforms). Each level handles specific types of distortions, allowing the system to progressively adapt to varying degrees of object transformation while maintaining accurate authentication
Solution Approach 2:
The system changes the parameters of the transform functions used for fingerprint alignment. By selecting appropriate transform types (similarity, affine, homographic) based on the detected distortion level, the system adapts its matching parameters to maintain accuracy across different viewing conditions and object states
2Measurement precision
If multiple transform types are used to handle complex distortions, then authentication accuracy improves, but the matching process becomes computationally intensive
Solution Approach 1:
The patent implements a dynamic, multi-level transform pyramiding process that adaptively selects the appropriate transform complexity. The system starts with simpler transforms and progressively applies more complex ones only when needed, optimizing the balance between matching accuracy and computational efficiency
Solution Approach 2:
The system applies transforms selectively based on the detected distortion level. Rather than always applying the most complex homographic transforms, the system uses partial action by applying only the necessary level of transform complexity for each specific matching case, reducing unnecessary computational overhead
3Reliability
If the system accounts for all possible changes in viewing angle and position, then authentication reliability improves, but the complexity of the matching algorithm increases
Solution Approach 1:
The patent segments the transform space into distinct levels (similarity, affine, homographic), each handling specific types of geometric transformations. This segmentation allows the system to account for various viewing angles and positions without requiring a single overly complex algorithm, instead using a structured hierarchy of simpler transform models
Data Source
AI summary
A system and method for better matching of two digital fingerprints when the two digital fingerprints are acquired under different conditions that uses the set of points of interest of the first and second digital fingerprints to perform the matching. A transform pyramiding process is performed using the first and second set of points of interest and a hierarchy of transforms to determine if the first and second set of points of interest are true matches despite the different conditions.