Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/usnistgov/NFIQ2/llms.txt

Use this file to discover all available pages before exploring further.

The FDA (Frequency Domain Analysis) module analyzes fingerprint ridge frequency patterns in the frequency domain to assess image quality. It computes histogram-based features that characterize the distribution of dominant frequencies across the fingerprint image.

​
Overview

Frequency Domain Analysis transforms image blocks into the frequency domain to measure ridge-valley structure clarity. The module divides the fingerprint image into blocks and analyzes the frequency characteristics of each block, creating a histogram distribution that represents the overall frequency patterns.
The FDA module is particularly effective at detecting blur and poor ridge-valley contrast, which manifest as anomalies in the frequency domain.

​
Class Definition

Header: quality_modules/FDA.h 
class FDA : public Algorithm {
public:
    FDA(const NFIQ2::FingerprintImageData &fingerprintImage);
    virtual ~FDA();
    
    std::string getName() const override;
    static std::vector<std::string> getNativeQualityMeasureIDs();
    static const char moduleName[];

private:
    std::unordered_map<std::string, double> computeFeatureData(
        const NFIQ2::FingerprintImageData &fingerprintImage);
    
    const int blocksize { Sizes::LocalRegionSquare };
    const double threshold { .1 };
    const int slantedBlockSizeX { Sizes::VerticallyAlignedLocalRegionWidth };
    const int slantedBlockSizeY { Sizes::VerticallyAlignedLocalRegionHeight };
    const bool padFlag { true };
};

​
Constructor

​
FDA()

FDA(const NFIQ2::FingerprintImageData &fingerprintImage)
Creates an FDA module instance and computes frequency domain features. Parameters:
  • fingerprintImage - Input fingerprint image data at 500 dpi
Throws:
  • NFIQ2::Exception - If the image resolution is not 500 dpi

​
Methods

​
getName()

std::string getName() const override
Returns the module identifier: "FrequencyDomainAnalysis" Returns: Module name as string

​
getNativeQualityMeasureIDs()

static std::vector<std::string> getNativeQualityMeasureIDs()
Returns the list of quality measure identifiers produced by this module. Returns: Vector containing feature IDs:
  • FDA_Bin10_0 through FDA_Bin10_9 - 10 histogram bins
  • FDA_Bin10_Mean - Mean of frequency values
  • FDA_Bin10_StdDev - Standard deviation of frequency values

​
Quality Features

The FDA module generates 12 quality measures:

​
Histogram Bins (0-9)

The fingerprint image is divided into blocks, and frequency domain analysis is performed on each block. The resulting frequency values are binned into 10 histogram categories based on predefined thresholds: 
static double FDAHISTLIMITS[9] = { 0.268, 0.304, 0.33, 0.355, 0.38, 0.407, 0.44, 0.50, 1 };
Each bin represents the proportion of blocks falling within a specific frequency range:
  • Lower bins (0-2): Low frequency content, indicating potential blur or poor clarity
  • Middle bins (3-6): Normal ridge frequency range for good quality fingerprints
  • Higher bins (7-9): High frequency content, may indicate noise or over-sharpening

​
Statistical Measures

  • FDA_Bin10_Mean: Average frequency value across all analyzed blocks
  • FDA_Bin10_StdDev: Standard deviation, measuring frequency consistency
The FDA module requires 500 dpi fingerprint images. Images at other resolutions will trigger an exception.

​
Configuration

The module uses the following default parameters:
ParameterValueDescription
blocksizeSizes::LocalRegionSquareSize of blocks for analysis
threshold0.1Minimum threshold for valid blocks
slantedBlockSizeXSizes::VerticallyAlignedLocalRegionWidthWidth for oriented blocks
slantedBlockSizeYSizes::VerticallyAlignedLocalRegionHeightHeight for oriented blocks
padFlagtrueEnable padding for rotated blocks

​
Usage Example

#include <quality_modules/FDA.h>

NFIQ2::FingerprintImageData fpImage = /* load image */;

try {
    NFIQ2::QualityMeasures::FDA fda(fpImage);
    
    // Get all feature IDs
    auto featureIDs = NFIQ2::QualityMeasures::FDA::getNativeQualityMeasureIDs();
    
    // Access computed features
    auto features = fda.getFeatures();
    double mean = features["FDA_Bin10_Mean"];
    double stddev = features["FDA_Bin10_StdDev"];
    
} catch (const NFIQ2::Exception &e) {
    // Handle error (e.g., wrong resolution)
}

​
Algorithm Details

The FDA algorithm:
  1. Divides the fingerprint image into blocks (typically 32x32 pixels)
  2. Applies Fast Fourier Transform (FFT) to each block
  3. Analyzes the dominant frequency components
  4. Computes frequency metrics aligned with ridge orientation
  5. Creates a histogram distribution of frequency values
  6. Calculates mean and standard deviation statistics

Build docs developers (and LLMs) love

Get started for freeTalk to us