Speaker identification using online, frame dependent, and diffusive variance adaptation | IEEE Conference Publication | IEEE Xplore