Biometrics Field Trial Evaluation Report

Section 5. Program Integrity—Usefulness of Biometrics in Strengthening Identity Management and in Detecting Fraud

5.1 Overview

The field trial demonstrated the capacity to fix client identity using biometrics and to increase confidence in decisions relating to client identity.

Field trial enrolments for visa applications totalled 14,854. Of those 14,854 enrolments, 394 matches were made because of multiple enrolments. Those match results show that biometric technology is a highly effective way to manage client identity:

• 97% of the fingerprint and facial biometrics enrolled were of high quality.
• When facial and fingerprint recognition were combined, the system made matches in 100% of cases.
• Verification was accurate in 96% of cases (see Section 7 for details).

In the 394 matches, the biometric search engine made a link to a previous field trial interaction. Of these, 12 matches were of particular interest from a program integrity perspective:

• One case was a clear case of fraud. The person had two separate applications under two different identities – one in the temporary visa program and one in the refugee protection program. The biometric system enabled CIC to make a link between the two identities, which would otherwise have been impossible.
• Two cases involved applicants who had been refused visas and who later reappeared as refugee protection claimants. The biometric system enabled CIC to make a link to previous visa data, which would normally have had to be done manually.
• In nine cases, the clients were issued visas and later claimed refugee protection. The biometric system enabled CIC to make a link between the refugee and visa data. This type of link, if there is full biometrics implementation, would enhance the decision‑making process.

5.2 Identity management

5.2.1 Managing client identities

Biometric technologies like the fingerprint and facial recognition technologies tested under the field trial have undergone significant industry testing to establish their effectiveness as identification tools. CIC wanted to test these technologies in day‑to‑day operations.

The field trial was designed to yield a statistically significant number of identification matches in order to test the quality and performance of biometric technology as an identity management tool.

CIC measured the following aspects of the performance of the field trial system:

• The quality of the collected biometrics, combined with performance measurements used in the industry
• The accuracy of matches with the number of repeat enrolments at the visa offices and at the Refugee Intake Centre
• The accuracy of matches with the number of clients who, once enrolled with a set of 10 fingerprints, enrolled one fingerprint at a port of entry for verification

Those key accuracy measurements gave CIC a better understanding of the benefits of biometric tools for its visa programs.

CIC manages a broad range of programs such as the temporary and permanent resident visa programs and the refugee protection program. Clients who apply under one program often appear later to change their status (for example, from visitor to student or from worker to permanent resident). As a result, CIC has many repeat clients. By providing an automated link to a previous application, biometric technology can help ensure that immigration officers have access to important case data, which can help them detect clients who try to obscure their immigration history by changing their name or date of birth and which can strengthen the level of trust between clients and CIC.

The field trial successfully tested the capacity of biometric technology to aid in managing the identity of repeat clients, in linking of case history and in detecting identity fraud.

5.2.2 Visa program integrity

Out of the 14,854 client enrolments in the field trial, 364 clients (2.5%) applied more than once. As a result, CIC had the following volumes of biometric samples for test purposes:

• 377 pairs of photos
• 212 pairs of fingerprints
• 195 pairs of photos and fingerprints together

Using the pairs of photos, testing showed that 98.4% of clients within this group were matched correctly using facial recognition technology alone; 97.9% were matched correctly using fingerprint technology alone; and 100% were matched correctly when fingerprint and facial recognition technologies were combined.

5.2.3 Entry management using a fingerprint biometric

The field trial also tested the capacity of fingerprint technology to verify clients on entry. Clients who had their 10 fingerprints enrolled at a visa office provided a single fingerprint at one of the participating ports of entry for comparison against their previously enrolled fingerprints.

During the field trial, of the 7,875 clients whose fingerprints had been enrolled at a visa office, 918 subsequently appeared at a field trial port of entry and had a single fingerprint verified. The field trial tested the capture of a verification fingerprint but did not send back a match or hit result in real time to the examining officer at the port of entry.

Verification was successful in 96.1% of cases, and no known cases of fraud were detected. In the other 3.9% of cases, the forensic specialists found the fingerprints to be of too poor quality to assess whether a match existed. Since no employees at port of entry sites participating in the field trial reported a client not matching their displayed visa photo, it seems credible that no fraud was attempted.

5.2.4 Identity management across the client continuum

During the field trial, the biometric samples of 14,854 visa applicants were compared against those of 3,410 refugee protection claimants to test whether clients had moved between programs. In 12 cases, visa applicants became refugee protection claimants during the six months of the field trial. Those 12 cases break down as follows:

• Nine cases involved individuals with valid visas presenting the same biographic data when they made their refugee protection claim. The biometric system enabled CIC to establish an automatic link back to the visa applications.
• Two cases involved individuals who had been refused visas and who had travelled without proper documentation to Canada to make a refugee protection claim. Again, the biometric system enabled CIC to trace these cases back to the initial visa application.
• One case involved a person making a refugee protection claim under another name and date of birth and concealing the fact that they had come to Canada with a visa. This instance was a clear case of identity fraud and shows that biometric tools are needed to prevent abuse of CIC’s programs. In this case, both facial recognition and fingerprint systems were highly accurate.

The 12 cases were discovered as a result of 13 biometric matches. In six cases, matches were made based on both facial and fingerprint data; in two cases, based on facial data alone (no fingerprints were enrolled for those clients at the field trial sites); and in five cases, on fingerprint data alone, because the system deemed that the client’s Refugee Intake Centre photos did not match those submitted at the visa office. For more information on matching and photo quality, see Section 6.

5.2.5 Biometrics as a deterrent

Other countries that have already implemented biometric systems have generally found they deter fraudulent activity.

Although hard to prove, there is some evidence to suggest that the field trial did deter visa recipients from arriving in Canada through the participating ports of entry. Before the field trial, client volumes for the same six‑month period in previous years were reviewed. In addition, a detailed survey of where visa recipients were expecting to land in Canada was conducted in Seattle. Based on those two analyses, without considering a possible deterrence factor, 71% of field trial clients were expected to arrive in Canada through either the Douglas or Pacific Highway land border crossings or through the Vancouver International Airport.

The purpose of that collection exercise was not only to plan the resources required for the field trial but also to have current comparison data at the end of the field trial to see whether the numbers dropped dramatically enough to conclude a deterrent effect.

Ultimately, only 10% of field trial clients arrived at one of the participating ports of entry, showing a marked difference between expected and actual arrivals.

One significant factor that may have changed travel patterns was the introduction of direct flights between Hong Kong and Toronto, which operated daily during the field trial. Clients who opted for this route would not have been verified in the field trial.

The sole case of identity fraud in the field trial involved a person claiming refugee protection at the Refugee Intake Centre in Toronto.

5.3 Conclusion

Biometric technology is an effective tool for confirming identity and detecting fraud. Matching performance was found to be high. Based on the accuracy of the results for clients known to have had multiple encounters with CIC, both fingerprint and facial biometrics performed well in identifying those clients.

<< Previous | Contents | Next >>