Quality Measures

Core functions for computing native quality measures and actionable quality feedback

The NFIQ2::QualityMeasures namespace provides functions to compute native quality measures and actionable quality feedback from fingerprint images.

Overview

Quality measures are individual metrics that assess specific aspects of fingerprint image quality, such as orientation certainty, local clarity, contrast, and minutiae characteristics. These measures can be computed independently or as part of the unified quality scoring process.

Compute Functions

These functions compute quality measures directly from fingerprint image data.

computeNativeQualityMeasureAlgorithms

std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>>
computeNativeQualityMeasureAlgorithms(
    const NFIQ2::FingerprintImageData &rawImage);

Compute native quality measures and return the evaluated algorithm objects.

rawImage · const NFIQ2::FingerprintImageData & · required

Fingerprint image in raw format. The image must be 500 PPI grayscale data, canonically encoded per ISO/IEC 39794-4:2019.

return · std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>>

A vector of evaluated native quality measure algorithms. Each algorithm object contains computed quality measure values and metadata.

Example:

#include <nfiq2_qualitymeasures.hpp>
#include <nfiq2_fingerprintimagedata.hpp>

NFIQ2::FingerprintImageData image(data, size, width, height, 0, 500);
auto algorithms = NFIQ2::QualityMeasures::computeNativeQualityMeasureAlgorithms(image);

---

computeNativeQualityMeasures

std::unordered_map<std::string, double>
computeNativeQualityMeasures(
    const NFIQ2::FingerprintImageData &rawImage);

Compute native quality measure values as a map of identifiers to values.

rawImage · const NFIQ2::FingerprintImageData & · required

Fingerprint image in raw format. The image must be 500 PPI grayscale data, canonically encoded per ISO/IEC 39794-4:2019.

return · std::unordered_map<std::string, double>

A map of quality measure identifiers to native quality measure values. Keys are identifiers from NFIQ2::Identifiers::QualityMeasures.

See also: Quality Measure Identifiers

Example:

auto measures = NFIQ2::QualityMeasures::computeNativeQualityMeasures(image);

for (const auto &[id, value] : measures) {
    std::cout << id << ": " << value << std::endl;
}

---

computeActionableQualityFeedback

std::unordered_map<std::string, double>
computeActionableQualityFeedback(
    const NFIQ2::FingerprintImageData &rawImage);

Compute actionable quality feedback values that indicate potential corrective measures for subsequent captures.

rawImage · const NFIQ2::FingerprintImageData & · required

Fingerprint image in raw format. The image must be 500 PPI grayscale data, canonically encoded per ISO/IEC 39794-4:2019.

return · std::unordered_map<std::string, double>

A map of actionable quality identifiers to actionable quality values. Keys are identifiers from NFIQ2::Identifiers::ActionableQualityFeedback. Values should be compared against thresholds in NFIQ2::Thresholds::ActionableQualityFeedback.

See also:

• Quality Measure Identifiers
• NFIQ2::Thresholds::ActionableQualityFeedback

Example:

auto feedback = NFIQ2::QualityMeasures::computeActionableQualityFeedback(image);

// Check if image has sufficient contrast
if (feedback["EmptyImageOrContrastTooLow"] > threshold) {
    std::cout << "Warning: Image may have insufficient contrast" << std::endl;
}

---

Extraction Functions

These functions extract quality measure values from already-computed algorithm objects.

getActionableQualityFeedback

std::unordered_map<std::string, double>
getActionableQualityFeedback(
    const std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>>
        &algorithms);

Obtain actionable quality feedback from computed quality measure algorithms.

algorithms · const std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>> & · required

Computed native quality measure algorithms, typically from computeNativeQualityMeasureAlgorithms().

return · std::unordered_map<std::string, double>

A map of actionable quality identifiers to actionable quality values. Keys are identifiers from NFIQ2::Identifiers::ActionableQualityFeedback.

Example:

auto algorithms = NFIQ2::QualityMeasures::computeNativeQualityMeasureAlgorithms(image);
auto feedback = NFIQ2::QualityMeasures::getActionableQualityFeedback(algorithms);

---

getNativeQualityMeasures

std::unordered_map<std::string, double>
getNativeQualityMeasures(
    const std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>>
        &algorithms);

Obtain native quality measures from computed native quality measure algorithms.

algorithms · const std::vector<std::shared_ptr<NFIQ2::QualityMeasures::Algorithm>> & · required

Computed native quality measure algorithms, typically from computeNativeQualityMeasureAlgorithms().

return · std::unordered_map<std::string, double>

A map of quality measure identifiers to quality measure values. Keys are identifiers from NFIQ2::Identifiers::QualityMeasures.

Example:

auto algorithms = NFIQ2::QualityMeasures::computeNativeQualityMeasureAlgorithms(image);
auto measures = NFIQ2::QualityMeasures::getNativeQualityMeasures(algorithms);

double orientationMean = measures["OrientationCertainty_Mean"];

---

Usage Patterns

Direct Computation

For simple use cases, compute quality measures directly:

NFIQ2::FingerprintImageData image(data, size, width, height, 0, 500);
auto measures = NFIQ2::QualityMeasures::computeNativeQualityMeasures(image);

Reusing Computed Algorithms

For efficiency when you need multiple outputs, compute algorithms once and extract values:

// Compute algorithms once
auto algorithms = NFIQ2::QualityMeasures::computeNativeQualityMeasureAlgorithms(image);

// Extract multiple outputs
auto measures = NFIQ2::QualityMeasures::getNativeQualityMeasures(algorithms);
auto feedback = NFIQ2::QualityMeasures::getActionableQualityFeedback(algorithms);
auto speeds = NFIQ2::QualityMeasures::getNativeQualityMeasureAlgorithmSpeeds(algorithms);

Performance Monitoring

Track computation time for each quality measure algorithm:

auto algorithms = NFIQ2::QualityMeasures::computeNativeQualityMeasureAlgorithms(image);
auto speeds = NFIQ2::QualityMeasures::getNativeQualityMeasureAlgorithmSpeeds(algorithms);

for (const auto &[alg, ms] : speeds) {
    std::cout << alg << " took " << ms << " ms" << std::endl;
}

Related

Quality Measure Algorithms

Learn about the Algorithm class and related functions

Quality Measure Identifiers

View all available quality measure and feedback identifiers