Feature-based speaker-independent word recognition without oral learning | IEEE Conference Publication | IEEE Xplore