Target-directed mixture dynamic models for spontaneous speech recognition | IEEE Journals & Magazine | IEEE Xplore