Neurotechnology identifies more than 6 million duplicate or inaccurate DRC voter records
More than five million duplicate voter registrations in the Democratic Republic of Congo have been identified by Neurotechnology in a massive multibiometric voter registration deduplication project.
The company announced the completion of the project less than two months after it began, with only 11 servers running its MegaMatcher ABIS turnkey solution to compare fingerprint and facial biometric records for 46.5 million voters. The system identified 5.3 million duplicates, and additionally more than 900,000 registered voters who do not meet age requirements. Human operators manually reviewed the records and confirmed the accuracy of the results.
“In the voter registry cleaning process our goal is to find as many duplicates and other noncompliant cases as possible,” said Mr. Corneille Nangaa Yobeluo, president of Commission Électorale Nationale Indépendante (CENI). “Neurotechnology helped us achieve our goals and exceeded our expectations by starting and completing the deduplication process in record time, and they were able to identify millions of duplicates.”
In conversation with Biometric Update at ID4Africa in Abuja, Nigeria, Neurotechnology Director Irmantas Naujikas told the publication that DRC law requires a human witness whenever a record is removed from the voter register. Two operators reviewed each case, or even three if the first two came to different judgements.
Removing duplicate records for more than 10 percent of the original voter base significantly enhances the likelihood of the upcoming DRC elections being completed successfully and without the results being challenged, but represented a major undertaking for both Neurotechnology and CENI. Neurotechnology Software Consultant Pavelas Gavlasevicius, who was on-site in the DRC, lauds the hard work of CENI in supporting project and reviewing with larger than expected number of duplicates.
“The biggest challenge for the project was the combination of the size of the database and the small amount of time in which it had to be completed,” says Naujikas. “They had to follow an exact deadline because of the election calendar, and it could not be changed, so it was very urgent. It was also a short amount of time for server hardware installation and testing.”
The company says that its MegaMatcher Accelerator Extreme (MMA Extreme), which the MegaMatcher ABIS is based on, is the fastest matching technology on the market. This speed allowed it to perform the deduplication on a country-wide scale, rather than dividing records by region or demographic criteria, which risks diminishing the accuracy of the results.
The combination of speed and accuracy is important to Neurotechnology, in terms of matching but also of tailoring its technology to the accommodate different kinds of identity systems. Business Development Manager Antonello Mincone says the technology is typically adapted and implemented in a couple of months.
Both the speed and accuracy of its algorithms have been demonstrated in NIST’s recently concluded MINEX III evaluation, which ranked Neurotechnology at or near the top of several categories.
NIST has also announced that Neurotechnology came first in speed and second in accuracy in the Iris Exchange (IREX) IX testing round. The company says its VeriEye and MegaMatcher solution was twice as fast as the sole entry which was found to be more accurate in terms of matching, and seven times faster at processing iris images.
“Once again the NIST IREX evaluation proved that our ongoing improvements in both accuracy and speed provide our customers with the best-performance solutions for real-time iris recognition biometric systems at any scale,” Neurotechnology’s biometric research head Dr. Justas Kranauskas said in the announcement.
For multi-biometric matching, Mincone says the design of Neurotechnology’s matching software is part of what makes it stand out from competitors.
“What we provide as a multi-biometric system, especially for large scale, is not just a combination of results, for instance based on Boolean decision making, but actually a true algorithmic fusion score that really works with the different biometrics together. That is important because it produces more accurate results. We are probably the only one offering that kind of multi-biometric system for matching from large scale databases at a very high speed.”
The company plans to pursue similar national ID projects at ID4Africa 2018, which runs April 24 to 26 in Abuja, Nigeria.
“To create, deploy and use Automated Biometric Identification Systems at a small scale is relatively easy; complexity and major challenges appear with scale,” said Denis Kacan, MegaMatcher ABIS product manager for Neurotechnology in the project results announcement. “With successful implementation of this project, our MegaMatcher ABIS and MegaMatcher Accelerator Extreme products proved to be suitable for large national scale deployments.”