Title :
Speaker-independent recognition of isolated Chinese digits by a pyramidical neural net
Author :
Yang, Shulin ; Ke, Youan ; Wang, Zhong
Author_Institution :
Dept. of Electron. Eng., Beijing Inst. of Technol., China
Abstract :
Neural networks have been widely applied to phoneme and word recognition in a speaker-dependent mode. In this paper, a new feature extraction approach, an extended multilayered neural net model and a modified learning algorithm were applied to speaker-independent recognition of isolated Chinese digits. Experiments show that about 85 percent accuracy can be obtained
Keywords :
learning systems; neural nets; speech recognition; extended multilayered neural net model; feature extraction; isolated Chinese digits; learning algorithm; phoneme recognition; pyramidical neural net; speaker-independent recognition; word recognition; Computer networks; Data mining; Fault tolerance; Feature extraction; Isolation technology; Multi-layer neural network; Neural networks; Robustness; Speech processing; Speech recognition;
Conference_Titel :
Circuits and Systems, 1991. Conference Proceedings, China., 1991 International Conference on
Conference_Location :
Shenzhen
DOI :
10.1109/CICCAS.1991.184276
