| Pulse coupled neural network (PCNN) is a kind of a background in biology, there are different from the traditional model of artificial neural network. There is a dual channel multiplication and pulse modulation for PCNN; and PCNN is a dynamic variable threshold of a new generation of artificial neural network. In recent years received widespread attention, scholars in the image has been widely applied in various research fields. Therefore, carries on the deep research has important theoretical significance and application value. In this paper, the neural network in image denoising and important application of speech recognition is studied.In the third chapter of this paper, first of all, through the median filtering, average filtering, wiener filtering and wavelet transform algorithm for image denoising, and highlight the main problems of the original algorithm existing in the process of image denoising. Next pulse coupled neural network is introduced the principle of image denoising algorithms, pulse coupled neural network model by MATLAB experiment was carried out on the digital image denoising and morphology operator, and compared with the traditional denoising algorithm can overcome the problems existing in the traditional algorithm, this algorithm can verify effectiveness and feasibility.In the fourth chapter, through to the speech signal feature extraction based on PCNN, then use the DWT algorithm and PNN algorithm for speech recognition, with the voice recognition technology based on MFCC and LPCC feature comparison, it is concluded that the extraction of the phonetic characteristics based on PCNN and target speech for speech recognition of the feasibility and validity of the algorithm. Experiments prove that the pulse neural network in the digital image denoising and speech recognition played a huge role.In the fifth chapter, all of this paper made a brief summary of the research content, and put forward the main research direction of next step. |
PDF Full Text Request