The Implementation And Optimization Of Speech Recognition Based On Neural Network On WinCE

Language is one of the most important ways to exchange information among the mankind. With the continuous development of computer technology and improvement of artificial intelligent, people expect to make the computer recognize human language. This requirement makes the technique of speech recognition have immense space to develop. Up to now, most speech recognition is based on conventional linear system theory, such as Dynamic Time Warping (DTW) and Hidden Markov Model (HMM). With the deep study of speech recognition, it is found that speech signal is a complex nonlinear process. To break through existing study, nonlinear system theory method must be introduced to it. With the development of artificial neural networks (ANN), it is possible to apply these nonlinear-system theories to speech recognition.This paper studied speech recognition based on ANN, computing validation, Performance analysis and results assessing are handled to each part of speech recognition process. A software platform was set up based on BP neural network model and a new way of mixing parameters of MFCC and LPCC was proposed. An application of speech verification was developed and optimized on WinCE which can recognize 0 to 9 at a rate of 88.4%.