Fingerprint Image Noise Removal via Local Texture Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing technologies face challenges in accurately identifying and removing sweat gland pores from fingerprint images, which are treated as noise, due to their similarity in appearance to bifurcation points and ridges, leading to inaccurate ridge extraction and varying noise levels across different areas of a fingerprint image.

Innovation Solution

An image processing apparatus and method that calculates local texture characteristics using LBP values and divides the image into local areas to identify specific characteristics, such as sweat gland pores and faded ridges, allowing for targeted noise reduction processing tailored to each area.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If sweat gland pores are treated as noise and removed using conventional methods, then noise reduction is achieved, but ridge extraction accuracy deteriorates because sweat gland pores are mistakenly identified as bifurcation points

Engineering Contradiction:

Improvenoise removal accuracyVSAvoidridge extraction accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The fingerprint image is divided into multiple local areas, and each local area is independently analyzed to determine its characteristic (ridge, sweat gland pore, or mixed). This segmentation allows for localized processing that preserves ridge structures while removing sweat gland pores, resolving the contradiction between noise removal and ridge extraction accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different processing approaches are applied to different local areas based on their identified characteristics. Areas with sweat gland pores undergo noise removal processing, while areas with ridges preserve their structural integrity. This local quality approach ensures that noise removal does not compromise ridge extraction accuracy in other areas.

Inventive Principle:

Principle #3Local quality

2Ease of operation

If uniform noise reduction processing is applied to the entire fingerprint image, then processing simplicity is maintained, but processing effectiveness deteriorates due to varying noise levels across different areas

Engineering Contradiction:

Improveprocessing simplicityVSAvoidnoise removal effectiveness

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The fingerprint image is divided into multiple local areas, each of which is independently evaluated for its characteristic. This segmentation enables the system to identify areas with varying noise levels and apply appropriate processing to each, thereby improving noise removal effectiveness while maintaining a relatively simple overall processing framework.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The processing approach dynamically adapts to the characteristics of each local area. The system automatically determines whether each local area contains ridges, sweat gland pores, or mixed characteristics, and adjusts the processing accordingly. This dynamic adaptation improves noise removal effectiveness without significantly complicating the processing workflow.

Inventive Principle:

Principle #15Dynamics

3Loss of information

If ridge extraction is performed before sweat gland pore recognition, then ridge structure information is available for analysis, but sweat gland pore identification accuracy deteriorates due to the chicken and egg problem

Engineering Contradiction:

Improveridge structure informationVSAvoidsweat gland pore identification accuracy

Core Design Contradiction:

Loss of informationVSMeasurement precision

Solution Approach 1:

The system performs preliminary analysis of each local area to determine its characteristic (ridge, sweat gland pore, or mixed) before performing ridge extraction or noise removal. This preliminary action resolves the chicken and egg problem by establishing the characteristic classification first, which then guides subsequent processing steps, ensuring both ridge structure information and sweat gland pore identification accuracy are preserved.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Instead of the conventional approach of extracting ridges first and then identifying sweat gland pores, the invention inverts the process by first determining the characteristic of each local area and then performing appropriate processing. This inversion resolves the circular dependency and enables accurate identification of both ridges and sweat gland pores.

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS10275677B2Image processing apparatus, image processing method and program

Publication Date: 2019.04.30 NEC SOLUTION INNOVATORS LTD

AI summary

An image processing device comprises an input unit, a characteristic amount calculation unit, a characteristic amount vector calculation unit, and a characteristic identification unit. The input unit receives an image. The characteristic amount calculation unit calculates an image characteristic amount characterizing a texture of a local area of the image received by the input unit. The characteristic amount vector calculation unit calculates a first characteristic amount vector corresponding to the local area from the image characteristic amount. The characteristic identification unit identifies a characteristic of the local area on the basis of the first characteristic amount vector and a second characteristic amount vector calculated by the same method as the first characteristic amount vector and calculated from an image whose characteristic has been determined in advance.