Overview
ISO/IEC 30116:2016 - "Information technology - Automatic identification and data capture techniques - Optical Character Recognition (OCR) quality testing" defines a standardized methodology for measuring and assessing the quality of OCR-B character strings. The standard specifies how to capture and process images of OCR symbols, how to quantify key attributes, and how to derive an overall quality grade. It also provides a reference decode algorithm for OCR-B and guidance on likely causes of deviations from optimum grades to support corrective action.
Key Topics and Technical Requirements
- Scope: Applies to OCR‑B (ISO 1073‑2) but the methodology can be applied partially or wholly to other OCR fonts.
- Image acquisition: Capture a high-resolution raw image under controlled, uniform illumination and best focus. Effective resolution should ensure character stroke widths span at least ten image pixels.
- Reference grey-scale image: Derived from the raw image by convolving pixel values with a synthetic circular aperture of 0.2 mm.
- Binarized image: Produced from the reference grey-scale using the thresholding algorithm defined in Annex B.
- Reflectance measurements: Calibrated reflectance values expressed as percentages (100% = barium sulphate or magnesium oxide reference). Use LED illumination at 890 nm and 940 nm; illumination elements ≤ 25 mm.
- Measurement parameters: Includes Best‑Fit (character placement), PCS/contrast, Position, Background Noise, Stroke Width Template (SWT) and Character Evaluation Value (CEV).
- Quality grading: Each parameter is graded as Recommended, Needs attention, or Not recommended. The lowest parameter grade determines the overall scan grade.
- Normative artifacts: Annex A (character centreline coordinates), Annex B (threshold determination), Annex C (reference decode algorithm), and worked examples for CEV calculations.
- Environmental/control notes: Ambient temperature guidance (20–25 °C) and IR absorption requirements for surrounding materials.
Applications and Who Uses It
- Identity document issuers and designers (passports, ID cards, driving licences) to validate OCR readability of MRZs.
- Test laboratories and quality assurance teams performing OCR quality testing and compliance verification.
- Manufacturers of document scanners, OCR engines and AIDC (automatic identification and data capture) devices for product validation and benchmarking.
- Border control and Automated Border Control (ABC) system integrators to ensure reliable MRZ recognition and secure access (e.g., BAC/SAC in e‑passports).
Related Standards
ISO/IEC 30116:2016 provides a practical, reproducible framework for assessing OCR quality, helping organizations improve OCR reliability and interoperability across document inspection and automated reading systems.