Fingerprint Recognition Using Eigenvector Template Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The large memory footprint of fingerprint templates due to storing numerous feature points of fingerprint images and the complexity of matching processes in fingerprint recognition systems, particularly when users wear face masks, necessitate a solution to reduce memory usage and improve user experience.
Innovation Solution
Perform dimensionality reduction processing on fingerprint feature data using a first orthogonal matrix generated through iterative calculation, converting feature vectors into eigenvectors to reduce memory occupation and simplify matching, employing Hamming distance calculation for improved efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fingerprint feature data is stored in traditional format, then recognition accuracy is maintained, but memory space occupied is large
Solution Approach 1:
The patent applies parameter changes by transforming fingerprint feature data from traditional high-dimensional representation to compressed low-dimensional representation through dimensionality reduction algorithms. This changes the parameter of data dimensionality while maintaining the essential recognition features, thereby reducing memory space occupation without significantly compromising recognition accuracy
Solution Approach 2:
The patent extracts the most critical feature components from the complete fingerprint feature set through dimensionality reduction. By identifying and retaining only the essential features that contribute to recognition accuracy while discarding redundant information, the system achieves compact storage with preserved functionality
2Reliability
If traditional matching methods are used, then matching accuracy is maintained, but matching complexity is high
Solution Approach 1:
The patent changes the parameter of feature representation from high-dimensional to low-dimensional space, which fundamentally simplifies the matching process. The reduced dimensionality decreases the computational complexity of comparison operations while preserving the discriminative power needed for accurate matching
Solution Approach 2:
The patent segments the fingerprint feature space into essential and non-essential components through dimensionality reduction. By separating and retaining only the critical features, the matching process operates on a simplified subset of features, reducing overall matching complexity while maintaining accuracy on the essential characteristics
3Reliability
If full feature data is used for matching, then recognition accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts and retains only the essential features that contribute most to recognition accuracy, discarding redundant information. This extraction process creates a compact feature representation that requires less processing time while maintaining the core discriminative capabilities needed for accurate recognition
Solution Approach 2:
The patent changes the time complexity parameter by reducing feature dimensionality. The transformed feature space enables faster processing operations while preserving the essential recognition information, achieving a favorable trade-off between processing speed and recognition accuracy
Data Source
AI summary
A method and an apparatus for fingerprint recognition are provided. The method is applied to the field of fingerprint recognition technologies. The method includes: acquiring a to-be-verified fingerprint image; determining a first eigenvector based on the to-be-verified fingerprint image, where the first eigenvector is obtained by performing dimensionality reduction processing on a second eigenvector using a first orthogonal matrix, and the second eigenvector is used to represent a feature of the to-be-verified fingerprint image; and performing fingerprint matching based on the first eigenvector and a third eigenvector, where the third eigenvector is obtained by performing dimensionality reduction processing on a fourth eigenvector using the first orthogonal matrix, the fourth eigenvector is used to represent a feature of a first fingerprint template, and the third eigenvector is stored in a fingerprint template library.