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 Mu module, also known as the Contrast module, measures the mean intensity and contrast characteristics of fingerprint images. It evaluates both global image contrast (Mu) and local block-level contrast uniformity (MMB).

​
Overview

The Mu module computes two key contrast-related features that are critical indicators of fingerprint image quality. Poor contrast can result from improper acquisition conditions, wet/dry fingers, or scanner limitations.
The module name “Mu” (μ) refers to the statistical mean. The module calculates the image mean and the mean of block means to characterize contrast distribution.

​
Class Definition

Header: quality_modules/Mu.h 
class Mu : public Algorithm {
public:
    Mu(const NFIQ2::FingerprintImageData &fingerprintImage);
    virtual ~Mu();
    
    std::string getName() const override;
    double getSigma() const;
    static std::vector<std::string> getNativeQualityMeasureIDs();

private:
    std::unordered_map<std::string, double> computeFeatureData(
        const NFIQ2::FingerprintImageData &fingerprintImage);
    
    bool sigmaComputed { false };
    double sigma {};
};

​
Constructor

​
Mu()

Mu(const NFIQ2::FingerprintImageData &fingerprintImage)
Creates a Mu module instance and computes contrast 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: "Contrast" Returns: Module name as string

​
getSigma()

double getSigma() const
Returns the computed sigma (standard deviation) value. Returns: Sigma value if computed Throws:
  • NFIQ2::Exception - If sigma has not yet been calculated

​
getNativeQualityMeasureIDs()

static std::vector<std::string> getNativeQualityMeasureIDs()
Returns the list of quality measure identifiers produced by this module. Returns: Vector containing:
  • Mu - Image mean (global contrast)
  • MMB - Mean of block means (local contrast uniformity)

​
Quality Features

The Mu module generates 2 quality measures:

​
Mu (Image Mean)

Feature ID: Mu The global mean intensity value of the entire fingerprint image. This measure indicates overall image brightness and contrast level. Interpretation:
  • Too low (< 100): Image may be too dark, poor contrast
  • Optimal range (100-150): Good contrast characteristics
  • Too high (> 150): Image may be too bright or washed out
Grayscale pixel values range from 0 (black) to 255 (white). An optimal fingerprint image typically has a mean around 127, indicating balanced ridge (dark) and valley (light) representation.

​
MMB (Mean of Block Means)

Feature ID: MMB The mean of all block-level mean values. The image is divided into blocks, the mean intensity of each block is calculated, and then the average of these block means is computed. Interpretation:
  • Similar to Mu: Indicates consistent contrast distribution
  • Significantly different from Mu: Suggests uneven illumination or contrast variations
Algorithm:
  1. Divide image into blocks (typically 32x32 pixels)
  2. Calculate mean intensity for each block
  3. Compute the average of all block means
// Simplified algorithm
const unsigned int blockSize = Sizes::LocalRegionSquare;
std::vector<double> vecMeans;

for (unsigned int i = 0; i < height; i += blockSize) {
    for (unsigned int j = 0; j < width; j += blockSize) {
        cv::Mat block = img(cv::Rect(j, i, takenBS_X, takenBS_Y));
        cv::Scalar m = mean(block);
        vecMeans.push_back(m.val[0]);
    }
}

// Calculate MMB
double mmb = 0.0;
for (auto mean : vecMeans) {
    mmb += mean / vecMeans.size();
}

​
Configuration

The module uses local region blocks defined by Sizes::LocalRegionSquare (typically 32x32 pixels) for computing the MMB feature.

​
Usage Example

#include <quality_modules/Mu.h>

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

try {
    NFIQ2::QualityMeasures::Mu mu(fpImage);
    
    // Get computed features
    auto features = mu.getFeatures();
    double imageMean = features["Mu"];
    double blockMeanAvg = features["MMB"];
    
    // Check contrast quality
    if (imageMean < 100 || imageMean > 150) {
        // Poor contrast detected
    }
    
    // Check contrast uniformity
    double contrastVariation = std::abs(imageMean - blockMeanAvg);
    if (contrastVariation > 10) {
        // Uneven contrast distribution
    }
    
    // Get sigma if available
    try {
        double sigma = mu.getSigma();
        // Use sigma value
    } catch (const NFIQ2::Exception &e) {
        // Sigma not computed
    }
    
} catch (const NFIQ2::Exception &e) {
    // Handle error
}

​
Use Cases

​
Detecting Acquisition Problems

  • Wet fingers: Often result in very low Mu values (dark image)
  • Dry fingers: Often result in very high Mu values (light image, poor ridge definition)
  • Dirty platen: May cause localized contrast variations (Mu vs MMB difference)
  • Poor sensor calibration: Consistent deviation from optimal range

​
Quality Assessment

The Mu module features are used in the NFIQ 2.0 quality score computation as indicators of:
  • Overall image contrast suitability for matching
  • Uniformity of contrast across the fingerprint area
  • Potential acquisition or sensor issues

Build docs developers (and LLMs) love

Get started for freeTalk to us