摘要：
Biometric-key computation is a feature binarization process, in which
a piece of live biometric data is transformed into a  bit-string with
auxiliary information. Biometric-key computation has been receiving an
increased amount of attention since biometric security schemes were
widely proposed. Emerging applications such as biometric-based
cryptographic key generation and biometric template protection require
biometric features  to be available in the binary form. This
representative string ought to be discriminative, informative and
privacy protective. However, it is commonly believed that satisfying
the first and the second criteria simultaneously is infeasible and a
tradeoff between them is always definite. In this talk, several recent
biometric-key generation schemes are reviewed. Specifically, we
present an efficient approach which involves discriminative feature
extraction, feature selection, quantization and Linearly Separable
Subcode-based encoding in fulfilling all properties mentioned above.

简历：
Andrew Teoh Beng Jin obtained his BEng (Electronic) in 1999 and Ph.D
degree in 2003 from National University of Malaysia. He is currently
an assistant professor in EE Department, College Engineering of Yonsei
University, South Korea. His research interest is in biometrics
security and pattern recognition. He had published around 180 refereed
international journal and conference papers in his area.