Line Noise Elimination via Projection Profile Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for eliminating line noise from latent fingerprint images often deteriorate the image quality and are ineffective when line noise lacks periodicity or has periodicity similar to fingerprint ridgelines, leading to misjudgment and incomplete extraction of fingerprint ridgelines.

Innovation Solution

A line noise eliminating apparatus and method that involves binarizing the image, rotating it by multiple angles to calculate edge features, determining line noise reliability, and applying local image enhancement to convert density without relying on periodicity, thereby accurately identifying and removing line noise in multiple directions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If Fourier transformation is used to eliminate line noise, then line noise can be removed, but the periodicity requirement limits effectiveness and fingerprint ridgelines may be eliminated when periodicity is similar

Engineering Contradiction:

Improveline noiseVSAvoideffectiveness against non-periodic or similar-periodicity noise

Core Design Contradiction:

Object-affected harmful factorsVSAdaptability or versatility

Solution Approach 1:

The patent changes the detection parameters from frequency-domain analysis (Fourier transformation) to spatial-domain analysis (projection profiles). By projecting pixel values along horizontal and vertical directions and analyzing the distribution of projection values, the method can detect line noise regardless of its periodicity, thus resolving the contradiction between noise removal effectiveness and adaptability to different noise patterns

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mathematical transformation mechanism (Fourier transformation) with a geometric projection mechanism. Instead of transforming the image to frequency domain and filtering, the method projects the image onto coordinate axes and analyzes projection value distributions, substituting a different mechanical/mathematical approach that is not constrained by periodicity requirements

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Object-affected harmful factors

If line noise elimination processing is applied, then line noise is removed, but the density of fingerprint ridgelines is deteriorated

Engineering Contradiction:

Improveline noiseVSAvoidfingerprint ridgeline density

Core Design Contradiction:

Object-affected harmful factorsVSManufacturing precision

Solution Approach 1:

The patent applies local quality by performing elimination processing only in regions identified as containing line noise, rather than applying global filtering to the entire image. By determining line noise regions through projection profile analysis and restricting processing to these specific areas, the method removes line noise while preserving the density and quality of fingerprint ridgelines in non-noise regions

Inventive Principle:

Principle #3Local quality

3Manufacturing precision

If related techniques are used to enhance fingerprint ridgelines, then ridgelines can be enhanced, but line noise cannot be properly distinguished when direction and periodicity cannot be extracted

Engineering Contradiction:

Improvefingerprint ridgeline enhancementVSAvoidline noise detection accuracy

Core Design Contradiction:

Manufacturing precisionVSMeasurement precision

Solution Approach 1:

The patent inverts the traditional approach by not trying to enhance ridgelines first and then remove noise, but rather by first identifying line noise regions through projection profile analysis and then applying enhancement only to non-noise regions. This inversion of the processing sequence allows accurate distinction between line noise and fingerprint ridgelines even when direction and periodicity are difficult to extract

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS7916957B2Line noise eliminating apparatus, line noise eliminating method, and line noise eliminating program

Publication Date: 2011.03.29 NEC CORP

AI summary

To provide a line noise eliminating apparatus and the like, with which the picture quality of the area that has no line noise is not deteriorated, and a line noise having no periodicity can be eliminated. The line noise eliminating apparatus includes: an image binarizing device which generates a binary image by an input image; a line noise reliability calculating device which calculates an edge feature quantity for each of black-pixel consecutive areas in the rotated images, and calculates line noise reliability based on the edge feature quantities; a line noise area determining device which determines the line noise areas that correspond to each of the rotation angle candidates based on the line noise reliability; a density converting device which generates a density-converted image by applying local image enhancement on an area that corresponds to the line noise area of the input image so as to generate a density-converted image.