This page is currently under development
Critical success factors


  1. Choice of biometric modalities

    1. Fingerprints: The most commonly used biometric. The technology was already far advanced, because of it's substantial contribution in providing forensic evidence in criminal investigations. Selection as one of the three biometrics in the ICAO document1 and strong support from law-enforcement agencies have contributed to its adoption for most national biometric ID projects. Because of the expense for the hardware required to perform matching, and the cost of the hardware needed for reliable data acquisition, choice of fingerprints is not suitable for systems that have budgetary constraints that result in compromising the data quality or the ability to perform the matching reliably. In most cases acquisition and authentication require contact with the equipment, and so quality degrades rapidly unless the equipment is cleaned frequently. Also, people who are employed in heavily labour intensive industries, particularly construction have almost no fingerprints.
    2. Iris recognition: Like fingerprints, iris recognition is one of the modalities recommended in the ICAO document1. John Daugman, of Cambridge University, almost single handedly promoted the technology. Bit his patents were often considered a constraint that drove decision makers away from making the technology more universallyly adopted, in spite of having lower error rates and requiring far less computing power for 1:N matching than fingerprints. In some cultures both acquisition and authentication are considered to be invasive. For example 30% of the female registrants in the Pakistani National ID project refused to enroll iris images.
    3. Facial: As the primary modality recommended in the ICAO document1, facial recognition initially enjoyed much success. The technology soon lost favor as its failings became evident. Even moderate disguises can fool facial recognition, and as a means of authentication, many simple "hacks" including using photographs of authorized personnel were found to circumvent the technology. It was also discovered that immigration officers were so focussed on ensuring that the finicky standards of acquisition took attention away from observation of behavioral traits that they had been trained to look out for. Facial recognition has lost much of its earlier standing.
    4. Vascular: The low error rates and undemanding processing requirements for vascular recognition are similar, if not better than iris recognition, without suffering any cultural resistance. Also, because acquisition and authentication is contactless, degradation of quality resulting from touching the equipment is not an issue. Because Japanese companies have led development in the field, it has only been the positive announcements of its successful use to prevent credit card fraud that has raised public awareness in Western Countries.
    5. Retina: Retina gained much of its positive reputation from science fiction novels and film. There are a number of similarities between iris and retina because both are ocular. The most significant difference is that iris uses a camera at close distance to acquire the data. Retina scanning requires a very close encounter with a scanning device that sends a beam of light deep inside the eye to capture an image of the Retina2. Because retina scanning is considered to be uncomfortable and intrusive, it has not been widely adopted.
    6. DNA: For forensic purposes, DNA is considered to be the most reliable biometric, with an estimated one in 10 billion chance of having a false match, far superior to any other biometric. The disadvantage is cost, with a single acquisition costing between $184 and $2403.
    7. Voice: As a means for forensic filtering, voice recognition allows technology to narrow copious data to allow technologists to focus on more manageable quantities of evidence. Voice recognition, like facial recognition does not have low enough error rates to allow stringent automated biometric processing.
    8. Handwriting: Handwriting, particularly in the form of the signature has been considered a reliable means of authentication. Now, as the science of biometrics has become increasingly automated, and machines are proving more effective that forensic handwriting experts, the relative superiority of other biometrics has relegated handwriting to the lower echelons. The other constraint that the biometric does not lend itself to rapid matching against large datasets makes it unsuitable for most ID systems.
    9. Ears: The acquisition of the ear for biometric is considered non-intrusive, and the shape of the ear, compared to facial features, changes relatively over a lifetime making superior to some biometrics. However, the error rates of existing technologies using ears are far higher than DNA, iris, retina and fingerprints, making it a less desirable choice than those other biometrics.
    10. Gait: The ability to analyze gait at a distance lends to the appeal of this biometric. As yet no reliable models have been developed to use gait, reinforced by the belief that gait can be readily disguised, and changes with aging and/or mobility.
    11. Hand geometry: This biometric technology is unsuitable for large identity systems because hand geometry is not unique to each individual. Also the size of each record is big, making storage and transmission an additional constraint.
    12. Palm print: Palm print recognition is a biometric technology which recognizes a person based on his/her palm print pattern. Palm print serves as a reliable human identifier because the print patterns are not duplicated in other people, even in monozygotic twins.

  2. Data quality

    1. Biometric data
      In many biometric applications, particularly when fingerprints are selected as the modality, data is collected manually. While this approach is considerably less expensive, unless the technicians acquiring the data have been thoroughly trained, and consistently apply the rigorous quality standards, sub-standard data acquisition results. This significantly affects the error rates, leading to duplicate registrations. For many applications de-duplication is one of the primary reasons for the adoption of the biometric standards.

      The issues experienced with duplicate voter registrations in Somaliland provide a vivid example of how poor quality of data acquisition can significantly dilute the ability of the system6.

    2. Demographics

      3. The personal information, including name, date of birth, gender and address form the foundation of the data for an identity system. Accurate recording of this information improves stakeholder support and facilitates matching data with other databases.

  3. Data contents

    1. Biometric and demographic data
    2. Links to existing data
    3. Address information
    4. Cell number
    5. Local landmark
    6. Card characteristics
    7. Card composition
    8. Polycarbonate
    9. Data content capability
    10. Encryption


  4. Card printing

    1. Laser
    2. Ink


  5. Delivery
  6. Buy-in from the public: There are a number of ways in which the public can benefit. In Malawi, a country without any credit bureaux, an experiment using fingerprints established that previously bad payers had improved payment trends once they could be reliably identified4. E-health, using biometrics is helping citizens in developing counties, mitigating fraud and streamlining payment5.

    Where biometrics make it easy to reliably authenticate voters ensure that disenfranchisement is kept to a bare minimum contributes to the credibility of elections and the legitimacy of the outcome7.
  7. Invested interest of government agencies
  8. Support at the most senior level of government
  9. Matching software
  10. Pilot studies
  11. Legal framework

  • Planning


    • More at:
    A Biometric Attendance Terminal and its Application to Health Programs in India
    A Cost-Benefit Analysis of Aadhaarz
    Banking the Poor via G2P Payments
    Biometric Technology in Machine Readable Travel Documents1
    Cash at Your Fingertips: Biometric Technology for Transfer Systems
    Democratic Governance, Citizenship, and Legal Identity
    Demographic and Health Surveillance of Mobile Pastoralists in Chad
    Dowa Emergency Cash Transfer (DECT) Wider Opportunities Evaluation and Recommendations
    E-health in low- and middle-income countries: findings from the Center for Health Market Innovations
    Ethiopia: The Revenue and Customs Authority to collect fingerprints
    Fingerprinting Individuals in the KEMRI/CDC Health and Demographic Surveillance System
    First Experiences in the Implementation of Biometric Technology to Link Data From Health and Demographic Surveillance Systems With Health Facility Data
    Guinea-Bissau: Second Poverty Reduction Strategy Paper
    Identification Strategy: A Field Experiment on Dynamic Incentives in Rural Credit Markets
    IHK introduces patient smart cards
    Immigrants stealing U.S. Social Security numbers for jobs, not profits
    India's Health Insurance Scheme for the Poor
    Innovative Pro-Poor Healthcare Financing and Delivery Models
    Inquiry Into the Procurement of Equipment for the National Security Information System (NSIS)
    Ministry to accept voter cards as valid ID
    Multiple problems thwart Nigeria SIM card registration
    Observation Mission of the Bolivia Voter Registration 2009
    Smartmatic assists Venezuela conduct national election
    Somaliland National Election Commission: Report on the Preparation of the Voter Register January to June, 2010
    The Identity Thief Killer

    Footnotes:
    1. This ICAO document establishes facial recognition supplemented by fingerprints and/or iris recognition as the biometric modalities of choice for International travel documents, including passports. It's influence has been profound. Many developed countries were influenced to adopt facial recognition and fingerprints as the modalities for travel documents and national ID.
    2. Iris Recognition vs. Retina Scanning – What are the Differences?
    3. Plex-ID Electrospray Time-of-Flight Mass Spectrometer for Mitochondrial DNA Base Composition Profiling pg 8
    4. Identification Strategy: A Field Experiment on Dynamic Incentives in Rural Credit Markets
    5. E-health in low- and middle-income countries: findings from the Center for Health Market Innovations pg 337
    6. Somaliland National Election Commission: Report on the Preparation of the Voter Register January to June, 2010 pg 1
    7. Smartmatic assists Venezuela conduct national election