Asynchronous integration of audio and visual sources in bi-modal automatic speech recognition | IEEE Conference Publication | IEEE Xplore