Fingerprint Image Orientation Accuracy via Weighted Gradient Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fingerprint image enhancement methods misjudge fingerprint orientation due to scars or defects on the finger, leading to inaccurate performance in identifying fingerprint orientation.
Innovation Solution
A fingerprint image enhancement method that uses a weighted image to increase confidence levels at ridge positions and decrease them at valley positions, allowing for accurate computation of fingerprint orientation without being influenced by defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional fingerprint image enhancement methods are used, then the enhancement process can be completed, but the fingerprint orientation determination is inaccurate due to scars or defects on the finger
Solution Approach 1:
The patent segments the fingerprint image processing by separating orientation determination from enhancement. It first determines orientation using gradient-based methods on the original image, then applies enhancement based on this pre-determined orientation, avoiding the interference of defects during orientation calculation
Solution Approach 2:
The patent performs preliminary orientation determination before enhancement. By calculating the orientation map from the gradient of the original fingerprint image before any enhancement operations, it ensures that defect-induced errors do not propagate into the enhancement process
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A fingerprint image enhancement method, comprising: receiving a fingerprint image; computing a horizontal variation image and a vertical variation image of the fingerprint image (300); computing a weighted image, wherein a weighted value of a first pixel corresponding to a finger ridge in the weighted image is greater than a weighted value of a second pixel corresponding to a finger valley (304); multiplying the horizontal variation image with the weighted image to generate a weighted horizontal variation image, and multiplying the vertical variation image with the weighted image to generate a weighted vertical variation image (306); computing a fingerprint orientation image according to the weighted horizontal variation image and the weighted vertical variation image (308); and performing fingerprint image enhancement on the fingerprint image according to the fingerprint orientation image (310).