A Method for Determining Illegal Broadcasting Based on Keyword
A technology for illegal broadcasting and keywords, applied in speech analysis, speech recognition, instruments, etc., can solve the problems of low keyword recognition rate and low recognition accuracy, so as to improve the accuracy of results, improve the error tolerance rate, and enhance specific recognition. The effect of the function
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0093] Such as figure 1 Shown, the present invention realizes by following technical scheme:
[0094] Step 1, designate the required keywords, set the detection threshold based on the word length, and save it to the keyword sequence list to establish the keyword sequence list;
[0095] Step 2, recording a broadcast audio file sample containing only keywords, and using the sample to train the acoustic model to obtain a mapping between the speech characteristics of the keyword and phonemes, and loading the mapping into the acoustic model to train the acoustic model;
[0096] Step 3, defining the mapping between phonemes and required specified words, and saving the mapping into the text phoneme sequence dictionary to establish a text phoneme sequence mapping dictionary;
[0097] Step 4, receiving broadcast audio data, and grouping the audio data according to the same frequency points in the same region;
[0098] Step 5, the broadcast audio data to be identified is roughly class...
Embodiment 2
[0146] The difference between this embodiment and embodiment 1 is that it also includes the following steps:
[0147] Step 7.7, compare the speech time domain sequence Yi(n)" or Xi(n)" obtained in step 7.6 with the speech time domain sequence Yi(n) or Xi(n) obtained in step 3, and calculate the residual sequence C i (m);
[0148] Step 7.8, for residual sequence C i (m) Execute the impulsive noise removal process in steps 8.1 to 8.5 to obtain a smooth residual sequence C i (m)";
[0149] Step 7.9, smoothing the residual sequence C i (m)" is compensated to the speech time-domain sequence Yi(n)" or Xi(n)" obtained in step 7.6 to obtain a new speech time-domain sequence Wi(n).
[0150] In the present embodiment, the specific method of setting up the SVM classifier in step ten is as follows:
[0151] Step 10.1, the Chinese character content is carried out statistics to the word of two characters, three characters, four characters, obtains the illegal word with higher word fre...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


