广播语音敏感关键词识别电子与通信工程专业论文.docxVIP

PAGE PAGE IV Broadcast Speech Sensitive Keywords Recognition Abstract Broadcast Speech Sensitive Keywords Recognition Abstract With the increasing number of interactive programs in broadcasting，the security of the broadcasting is also facing a major challenge. Compared to traditional manual means of monitoring，automatic audio processing technology due to its advantage is gradually applied to the monitoring and management of the broadcast content. According to the characteristics of the broadcast program content monitoring，the technology of voice keyword recognition (KWR) is a suitable choice for the automatic monitoring of sensitive information in the interactive broadcast program. The technology of voice keyword recognition (KWR) can be realized in different ways. In this thesis，taking HTK as a tool， in MATLAB environment，based on hidden Markov model (HMM) ，according to the principle of using HTK in isolated word recognition system，a means of voice keyword recognition (KWR) in the interactive broadcast program，especially the hotline telephone interview program was achieved. In order to better targeted，in this thesis the voice data used for training and testing are all from the real recording of the broadcast program，to be used in the testing of the keyword recognition (KWR) system. The main work of this thesis is as follows: 1 ， This thesis briefly introduces the current situation of speech recognition technology，especially the application in broadcasting and television. Further discusses the effect of keyword recognition (KWR) technology used for the automatic monitoring of sensitive information in the interactive broadcast program ，gives related background materials and the purpose of this research. 2，The basic principle and structure of the voice keyword recognition system are explained，and the keyword recognition method based on keywords and non- keywords (fillers) network is decided to select as a way of speech recognition for