Rotation-Invariant Fingerprint Recognition via Ridge Flow Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fingerprint recognition systems face difficulties in comparing fingerprint images oriented at different angles, leading to increased processing effort and user frustration, which can hinder the practicality of fingerprint recognition security.
Innovation Solution
A fingerprint recognition sensor processes fingerprint image data to generate a rotation-invariant representation, using directional vectors, differences of directions, and Poincare index values, allowing for efficient comparison of fingerprints regardless of orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If fingerprint images are compared directly without rotation normalization, then processing time is reduced, but comparison accuracy deteriorates when fingerprints are oriented at different angles
Solution Approach 1:
The system performs preliminary extraction of rotation-invariant features (minutiae points, ridge flow patterns, Poincare indices) during the enrollment phase and stores them in a normalized form. During authentication, these pre-extracted features are directly compared without requiring time-consuming rotation normalization operations, thus resolving the contradiction between processing speed and accuracy.
Solution Approach 2:
The invention extracts and separates the rotation-invariant characteristics from the fingerprint images, such as minutiae locations, ridge flow directions, and Poincare indices. By taking out these invariant features and comparing them independently from the original image orientation, the system achieves accurate comparison without the computational burden of full image rotation normalization.
2Measurement precision
If rotation normalization is applied to fingerprint images, then comparison accuracy at different angles is improved, but processing effort and time increase substantially
Solution Approach 1:
The system replaces the mechanical operation of rotating and normalizing fingerprint images with a mathematical approach using Poincare indices and ridge flow analysis. This substitution eliminates the need for actual image rotation while achieving the same normalization effect, thereby maintaining accuracy without the computational overhead of traditional rotation methods.
Solution Approach 2:
The invention transforms the fingerprint representation from pixel-based images to feature-based parameters including Poincare indices, ridge flow directions, and minutiae characteristics. These parameter transformations are rotation-invariant by nature, allowing direct comparison of fingerprints at different orientations without performing time-consuming rotation operations.
3Measurement precision
If multiple rotation angles are tested through repeated operations, then comparison accuracy is improved, but device complexity and processing effort increase
Solution Approach 1:
The system segments the fingerprint comparison task into independent rotation-invariant feature extraction and comparison modules. By dividing the fingerprint into distinct features (minutiae, ridges, Poincare indices), each can be processed and compared independently without requiring complex coordination between different rotation angles, thus reducing overall system complexity.
Solution Approach 2:
The invention introduces Poincare indices and ridge flow patterns as intermediary representations that mediate between the original fingerprint images and the comparison process. These intermediaries capture the essential fingerprint characteristics in a rotation-invariant form, eliminating the need for complex multi-angle rotation operations while maintaining comparison accuracy.
Data Source
AI summary
A method includes receiving fingerprint image data at a fingerprint recognition sensor, where the fingerprint image data are associated with an authorized user. The fingerprint image data are transformed into a substantially rotationally invariant representation, which is maintained in a database of enrolled fingerprint information. Processed fingerprint image data from an accessing user are compared with the substantially rotationally invariant representation of the fingerprint image data from the authorized user.