A team of engineers at the University of Ilorin has produced a prototype biometric machine that is capable of eliminating a major deficiency encountered by imported machines.
The team leader, Prof. Tunji Samuel Ibiyemi of the Electrical and Electronics Engineering Department, said that with the use of local content, the researchers have been able to unravel the challenge of the inability of the imported biometric machine to adequately recognise physical features of black people.
Prof. Ibiyemi disclosed that imported solutions to the challenges of impersonation, economic fraud, multiple voting, examination malpractices, election rigging, and security challenge do not work optimally among black people as they do among white people, adding that “what we have just produced one that works better for blacks.”
The don, whose research effort was sponsored by the World Bank-assisted Science and Technology Education Post-Basic (Step-B) Project, explained that “this lack of local content in the making of the machines we use for vital national assignments perhaps explain why government efforts on projects like e-voting, national identity card scheme, security intelligence on criminal citizens had not been yielding enough fruits.”
Prof. Ibiyemi led a team of three other researchers, Prof. J. Sadiku of Computer Science Department, Dr. S. A. Aliu and Dr. I. O. Avazi of Electrical and Electronics Engineering Department, to study ‘Biometric Signal Processing for Personal Application and Forensic Application’.
The Professor of Electrical and Electronics Engineering said, “An average Chinese recognises Chinese people more easily. Likewise, Americans know one another better. When I was in Britain, any black person could pick any of his friend’s identity card and go anywhere across the country unfettered. Those working at the airports will confirm to you that you need local people to identify one another.
“Most imported machines don’t recognise tribal marks. These machines raise alarm when they see a masked face. But what we have produced can recognise tribal marks, faces that are masked and faces that are disguised using cosmetics.”
Prof. Ibiyemi pointed out that the machines developed by his team are not only more cost-effective than their imported equivalents, they are also more functionally efficient. He said, “For instance, after our proposal was approved for funding and we were experiencing delays in accessing the money, we funded the development of a locally made Iris Scanner at N60, 000. But when we eventually got money to buy this equipment, we got it for N400, 000 (that is 2,500 US dollars). And the former produces sharper and clearer images than the latter.”
He explained further, “Besides this, we used machines (computers) to recognise the human face, human iris, finger prints, toe prints and sole prints. We worked on speaker and speech recognition, signature verification and hand writing verification.
“What we found is amazing. For example, using any of these parameters, we can get different patterns of iris, fingerprints, toeprints, soleprints that differentiate the 14 billion people in the world. No two persons have the same pattern for any of these parameters. In fact, the left iris pattern for an individual is different from the right side for the same individual. Similarly, the thumb print is different from the index fingerprint for each person and so on.
“No website is available within Africa for biometric data on black people. Foreign data are populated by white people. We needed black people’s data to work with, so we started our own website-www.unilorin.edu.ng/step-b/biometrics.
“We collected over one million fingerprints, 600,000 toeprints, 200 soleprints, and 374 latent fingerprints. The website is hosted in Italy. We also went to the home of lepers and took 200 soleprints. When we brought them to the laboratory, we discovered that it is easier to recognise people through their soleprints than through their fingerprints.
“Our fingerprint equipment will identify each individual. We can detect multiple voting. There are five groupings of finger prints. Using one million samples, we compared our result with US-based Federal Bureau of Intelligence (FBI) result,” he said.