This paper gives a status report of the Philips research system for phoneme-based, large-vocabulary, continuous-speech recognition. Like for many other systems, the recognition architecture is based on an integrated statistical approach. We describe the characteristic features of the system as opposed to other systems: 1. The Viterbi criterion is consistently applied both in training and testing. 2. Continuous mixture densities are used without tying or smoothing. 3. Time-synchronous beam search in connection with a phoneme look-ahead is applied to a tree-organized lexicon. The system has been successfully applied to the American English DARPA RM task. Here, we report experimental results for a German 13 000-word Philips internal dictation task. In addition to the scientific prototype, a PC version has been set up which is described here for the first time.
Keywords: Continuous speech recognition, large vocabulary recognition
Bibliographic reference. Steinbiss, Volker / Ney, Hermann / Haeb-Umbach, Reinhold / Iran, B.-H. / Essen, U. / Kneser, Reinhard / Oerder, M. / Meier, H.-G. / Aubert, X. / Dugast, Christian / Geller, D. / Hollerbauer, W. / Bartosik, H. (1993): "The Philips research system for large-vocabulary continuous-speech recognition", In EUROSPEECH'93, 2125-2128.