Speech Recognition Algorithm Based on Nonlinear Partition and GFCC Features

In order to speed up and enhance the robustness of speech recognition system, this paper proposes a speech recognition algorithm based on segment-level features of GFCC. In training and testing stage we use segment-level features of GFCC which is more robust to noise instead of the widely used MFCC features. Experiment results show that both the training time and test time decreased, while the accuracy of system was made to improve.

[1]  Thomas Fang Zheng,et al.  Segmentation-based Method for Text-Dependent Speaker Recognition in Embedded Applications , 2010 .

[2]  Alex Acero,et al.  Spoken Language Processing , 2001 .

[3]  Hongzhi Wang,et al.  Study on the MFCC similarity-based voice activity detection algorithm , 2011, 2011 2nd International Conference on Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC).

[4]  Yi Jiang,et al.  Auditory features based on Gammatone filters for robust speech recognition , 2013, 2013 IEEE International Symposium on Circuits and Systems (ISCAS2013).