07 February 2018
Associate Professor Terence Sim began his research in facial recognition technology since 1998. He said that while facial recognition can be used in many fields, but it is used mainly for security purposes due to the ease of using cameras to carry out large scale screening. However, A/P Sim highlighted that the rate of error for facial recognition is higher than fingerprint and iris recognition. He cautions that the authories be careful before adopting the technology due to its high risk. A/P Sim recommends for the government to create a more comprehensive law to protect privacy to prevent identity theft. Technology companies need to ensure that their data is secure and materials are not easily stolen. A/P Sim also highlighted the necessity to distinguish between "proof of identity" and "proof of authorisation". One pitfall of the technology is that the technology is not transferrable and cannot be cancelled, unlike passwords and physical passes. NUS Computing is studying how to make facial recognition more secure and it is developing a dynamic face recognition system to improve accuracy. A trial on 39 pairs of twins proved successful, and A/P Sim said it is because everyone's expressions are different and not easily replicated.
Lianhe Zaobao, 7 February 2018