| H93-1023 | the top 16 IMELDA parameters for | speaker-independent recognition | . A different IMELDA transform |
| H91-1004 | year , special projects studying | speaker-independent recognition | based on stored phoneme prototypes |
| H90-1075 | began in June 1989 . The first | speaker-independent recognition | result , 86 % , was obtained |
| H91-1080 | future developments in improved | speaker-independent recognition | . Developed a new formalism for |
| H90-1079 | future developments in improved | speaker-independent recognition | . <title> SPOKEN LANGUAGE SYSTEMS |
| H90-1075 | neural network classification to | speaker-independent recognition | of spoken letters . The first |
| H92-1100 | develop improved acoustic models for | speaker-independent recognition | of continuous speech , together |
| H91-1083 | develop improved acoustic models for | speaker-independent recognition | of continuous speech , together |
| H89-2064 | been exploring techniques for | speaker-independent recognition | . Various avenues have been explored |
| H89-1014 | been exploring techniques for | speaker-independent recognition | . Various avenues were explored |
| H89-2032 | the rapid configuration of new | speaker-independent recognition | tasks , incorporating new lexical |
| H90-1082 | develop improved acoustic models for | speaker-independent recognition | of continuous speech , together |
| H89-2062 | develop improved acoustic models for | speaker-independent recognition | of continuous speech , together |
| H94-1038 | vocabulary , sec \ -LSB- 3 \ -RSB- . A | speaker-independent recognition | system has been built according |
| H90-1075 | multivariate classifiers to obtain 89 % | speaker-independent recognition | of spoken letters \ -LSB- 5 \ |
| H92-1055 | has been an early proponent of | speaker-independent recognition | . While most of our work presented |
| H91-1053 | adaptation of types 1 and 2 . | Speaker-independent recognition | results are also shown for comparison |
| H92-1077 | efforts to develop high-performance | speaker-independent recognition | techniques . Next , Fil Alleva |
| H90-1049 | intervals . Other work includes " | Speaker-Independent Recognition | of Connected Utterances Using |
| H91-1009 | models are not good enough to do | speaker-independent recognition | , but they serve as a better |
