Fingerprint Image Segmentation for Real-Time Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fingerprint images acquired by direct view sensors, such as those based on thin-film transistor technology, suffer from low contrast and parasitic zones like shadows, making them incompatible with conventional authentication systems, and existing processing methods are not suitable for real-time processing of high-resolution images acquired at rates over 10 frames per second.
Innovation Solution
A method that segments fingerprint images by assigning frequency response levels to pixels based on gray level variability, grouping neighboring pixels with similar frequencies, and selecting regions with frequencies above a threshold to generate a modified image suitable for authentication, which includes steps like morphological erosion, dilation, and median filtering to enhance contrast and remove irrelevant areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If fingerprint images are acquired by direct view sensors, then acquisition speed is improved (over 10 frames per second), but image quality deteriorates (low contrast and parasitic zones)
Solution Approach 1:
The patent applies segmentation by dividing the fingerprint image into multiple frequency bands using wavelet transform. This allows separation of the fingerprint signal from low-frequency background elements like shadows and parasitic zones, thereby improving image quality while maintaining the high acquisition speed of direct view sensors
Solution Approach 2:
The patent changes the frequency domain parameters of the image through wavelet transform and selective frequency band processing. By transforming from spatial domain to frequency domain and selectively enhancing certain frequency components, the image quality is improved without compromising the acquisition speed
2Measurement precision
If conventional latent fingerprint processing methods are used, then image enhancement is improved, but processing speed deteriorates (not suitable for real-time processing)
Solution Approach 1:
The patent applies preliminary action by performing wavelet transform and frequency band selection as preprocessing steps before authentication. This preliminary enhancement of the image in the frequency domain prepares the data for faster subsequent processing, enabling real-time authentication while maintaining high image quality
Solution Approach 2:
The patent substitutes traditional spatial-domain image processing methods with frequency-domain processing using wavelet transform. This mathematical transformation approach replaces conventional mechanical or iterative enhancement methods, providing both high image quality and computational efficiency for real-time processing
3Measurement precision
If high-resolution images are processed, then authentication accuracy is improved, but processing time increases (exceeds real-time requirements)
Solution Approach 1:
The patent applies the extraction principle by isolating and retaining only the relevant frequency components that contain fingerprint information. By extracting and discarding irrelevant low-frequency components (shadows, parasitic zones), the processing load is reduced while maintaining authentication accuracy, enabling real-time processing of high-resolution images
Data Source
Figure 1
Figure 2~3b
Figure 3c
AI summary
The invention relates to a method for processing an image comprising a set of pixels, each pixel being associated with a level of gray, the method comprising an image segmentation step to generate a modified image containing only regions of the image exhibiting an alternation of light and dark areas at a frequency greater than a minimum frequency, said segmentation step comprising: ∘ the assignment, to each pixel of the image, of a frequency response level, corresponding to a frequency of alternations of light and dark areas in the vicinity of the pixel, ∘ the definition of image regions by grouping neighboring pixels of the same frequency response level, ∘ the determination of a threshold frequency response level, and ∘ the generation of an image comprising only regions whose pixels exhibit a frequency response level greater than or equal to the threshold frequency response level.