Modeling method based on residual convolutional network, voice recognition method and electronic equipment
A technology of convolutional network and modeling method, applied in modeling method based on residual convolutional network, speech recognition method and electronic equipment field, can solve the problem of poor accuracy, inability to use context information, acoustic model audio signal representation Inefficiency and other problems to achieve the effect of improving accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0052] see figure 1 , the embodiment of the present invention provides a residual convolutional network-based modeling method that can be applied to electronic equipment. The electronic equipment can be, but not limited to, computers, mobile phones, or televisions. When the method is applied to electronic equipment, the execution Step S110 to step S140:
[0053] Step S110: Obtain a plurality of audio data in the corpus.
[0054] Wherein, the corpus is a large-scale electronic text library that has been scientifically sampled and processed, and has a large amount of speech data inside. The speech data may include but not limited to Chinese speech data, English speech data or other foreign language speech data. It can be understood that, The corpus may also include vocabulary phrases and article paragraphs.
[0055] In this embodiment, the corpus is a Chinese corpus, wherein the Chinese corpus may include THCHS30 Chinese voice data set, ST-CMDS Chinese voice data set, AISHELL ...
Embodiment 2
[0100] see Figure 4 , on the basis of the first embodiment above, the present invention also provides a voice recognition method applicable to electronic equipment, the electronic equipment can be but not limited to computers, mobile phones or televisions, etc., when the method is applied to electronic equipment, When performing speech recognition, the acoustic model generated in the first embodiment above is used for speech recognition, and steps S210 to S230 are executed.
[0101] Since the speech recognition method uses the acoustic model in the above-mentioned first embodiment, for the specific description of the acoustic model, reference may be made to the specific description of the above-mentioned first embodiment, and details will not be repeated here.
[0102] Step S210: Obtain voice data to be recognized.
[0103] The way to obtain the speech data to be recognized can be to obtain the speech data collected by the speech recognition module, or to obtain the speech d...
Embodiment 3
[0112] The present invention also provides an electronic device, the electronic device includes a memory and a processor, and a computer program is stored on the memory, and when the computer program is executed by the processor, the above-mentioned modeling based on the residual convolution network is performed method or voice recognition method.
[0113] The memory and the processor are directly or indirectly electrically connected to each other to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More