Calculation of Quality Score

[Pragya279]

User Story: Calculation of Quality Score

As a Registration Operator
I want the system to automatically calculate and display the quality score for each biometric modality captured
So that I can ensure the biometrics meet the minimum quality requirements before submitting the registration packet.

Problem / Need

Biometric capture quality varies based on lighting, sensor conditions, operator skill, and applicant posture. Poor-quality biometrics lead to high false match rates, delays in identity creation, and downstream failure in authentication or deduplication.

To improve data quality and ensure compliance with biometric standards (e.g., ISO/IEC 29794), the system must calculate, display, and enforce quality thresholds for fingerprints, face, and iris captures.

Feature Description

Automatic Quality Calculation
After capturing any biometric (fingerprint, iris, face), the system automatically triggers quality evaluation.

Quality score is calculated either by:

An internal MOSIP quality algorithm, or

A configured external third-party quality SDK (e.g., NFIQ2).

The calculation is real-time, with no manual intervention required.

Score & Status Output
The output includes:

Numerical score (e.g., 0–100 or SDK-specific range)

Pass/Fail status based on threshold

Optional granular metrics (contrast, sharpness, clarity, noise, focus, pose estimation)

Display to Operator
The UI shows:

A visual indicator (green/yellow/red)

The exact score

Pass/Fail message

If the quality is below threshold, a prompt instructs the operator to recapture.

Configurable Thresholds
Each biometric modality has its own configurable minimum score:

fingerprint.threshold

iris.threshold

face.threshold

Threshold values are controlled via configuration files or server-side policies.

Enforcement Logic
If score ≥ threshold → operator can proceed

If score < threshold → operator must recapture (configurable to allow override for exceptions)

Data Storage
The final quality score and status are stored as part of the biometric packet metadata.

Supports future audits and analytics.

Error Handling
If the quality calculation fails due to SDK error, device failure, or invalid inputs:

The system shows a clear error message

Allows retry or fallback to default device quality metrics

Acceptance Criteria
AC1: Automatic Trigger

Quality score calculation must start immediately after each biometric capture.

AC2: Valid Score Generation

The system must generate a numerical score and pass/fail decision.

AC3: Operator Visibility

The UI must clearly show the calculated score and its meaning.

AC4: Threshold Enforcement

Operators cannot proceed with low-quality biometrics unless override is enabled in configuration.

AC5: Modality Coverage

Quality scoring is supported for:

All fingerprint types (slap, thumb, rolled if applicable)

Iris captures

Face images

AC6: Storage

The system stores the quality score in the biometric packet.

AC7: Configurability

Administrators can modify thresholds without code changes.

AC8: Error Handling

Appropriate error messages and fallback logic must be provided.

Example User Flow

Operator captures a face image.

System calculates score: 62, threshold = 70 → FAIL

UI shows a red indicator and message:
“Face quality too low. Please recapture.”

Operator recaptures; new score = 78 → PASS

Operator proceeds to next step.