A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2003; you can also visit the original URL.
The file type is
Face recognition has attracted much attention in the past decades for its wide potential applications. Much progress has been made in the past few years. However, specialized evaluation of the state-of-the-art of both academic algorithms and commercial systems illustrates that the performance of most current recognition technologies degrades significantly due to the variations of illumination and/or pose. To solve these problems, providing multiple training samples to the recognition system isdoi:10.1109/icme.2003.1221270 dblp:conf/icmcs/GaoSCF03 fatcat:zpa643v2pfg67f3bfxqqeplp4i