Fingerprint Image Segmentation for Real-Time Authentication

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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)

Engineering Contradiction:
Improveacquisition speedVSAvoidimage quality
Core Design Contradiction:
SpeedVSMeasurement precision

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

Inventive Principle:
Principle #1Segmentation

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

Inventive Principle:
Principle #35Parameter changes

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)

Engineering Contradiction:
Improveimage enhancementVSAvoidprocessing speed
Core Design Contradiction:
Measurement precisionVSProductivity

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

Inventive Principle:
Principle #10Preliminary action

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

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If high-resolution images are processed, then authentication accuracy is improved, but processing time increases (exceeds real-time requirements)

Engineering Contradiction:
Improveauthentication accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

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

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3214601B1Image enhancement method applicable to fingerprint images
Publication Date: 2020.10.07 IDEMIA IDENTITY & SECURITY FRANCE SAS
  • EP3214601B1 patent drawingFigure 1
  • EP3214601B1 patent drawingFigure 2~3b
  • EP3214601B1 patent drawingFigure 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.