Portia Khanyile is speeding up both the processing of fingerprint images and the searching of databases.
Young, bright and equipped for the job, the CSIR's Portia Khanyile is making inroads into the speeding up of identification processes. This requires improving the speed of both the processing of fingerprint images and the searching of databases. "Fingerprint identification is computationally intensive and requires a lot of resources such as processing power," Khanyile summarises the challenge.
Public and private institutions' databases are becoming larger on a daily basis. This is affecting the database search speed - and not ideal for industries such as banking. The swelling of databases is largely attributed to the move from serial biometric systems towards parallel solutions which include the use of Graphics Processing Unit and Field-Programmable Gates.
Illustrating the issue at stake, she says: "Think of bigger databases such as the Integrated Automation Fingerprint Identification System, which has about 64 million criminal fingerprint images. It takes about 10 minutes to search this database. A response time of 10 minutes may be acceptable for forensic investigations, but for commercial purposes, it is simply not ideal."
Khanyile is one of a group of students placed
on a studentship programme implemented at CSIR Modelling and
Digital Science. At the end of August, she will be submitting
her Master's dissertation titled, Fingerprint identification
using distributed computing.
High performance computing holds the key to the required improvement in speed, and Khanyile used the Centre for High Performance Computing in Cape Town to conduct her research and get closer to a solution. This work has resulted in three conference publications and a journal paper.
"This is a great achievement for me considering the fact that I am about to submit my dissertation, way before time," she says.
Khanyile completed her undergraduate degree in computer science and statistics at the University of KwaZulu-Natal. Her Honours degree was in computer science at the same university. She is also an Institute of Electrical and Electronics Engineers certified biometric professional.
She hails from Dundee in northern KwaZulu-Natal.