Method and device for recognizing state of meridian

1. A method for recognizing a state of a meridian, wherein the state of the meridian is used to indicate a functional status of the meridian, comprising:

receiving an input voice of a user;

preprocessing the input voice;

extracting a stable feature of the preprocessed input voice;

primarily classifying the stable feature on the basis of a feature recognition model, and determining a basic classification pitch, wherein the basic classification pitch comprises Gong, Shang, Jue, Zhi and Yu (respectively equivalent to do, re, mi, sol and la);

secondarily classifying the stable feature on the basis of the feature recognition model, and determining a secondary classification tone in the basic classification pitch; and

recognizing the state of the meridian according to the secondary classification tone,

wherein, said extracting a stable feature of the preprocessed input voice comprises:

transforming a sequence of the input voice into a frequency domain feature through Fourier transform to obtain a signal power spectrum, wherein the sequence of the input voice includes time and spatial sequences of the input voice;

utilizing discrete cosine transform to remove a correlation between the signal dimensions of the signal to obtain a 24-dimensional feature of a sample, wherein the signal dimensions refer to intervals having different frequencies; and

utilizing a manifold learning method to reduce the 24-dimensional feature to be a 12-dimensional feature to obtain the stable feature,

wherein the step of receiving an input voice of a user comprises:

sequentially displaying first to fifth prompt words and starting to receive first to fifth input voices of a user;

after the first to fifth input voices of the user are received, recognizing the received first to fifth input voices to be first to fifth user words;

respectively comparing the first to fifth user words with the first to fifth prompt words; and

integrating the user words matched with the prompt words to be an audio file as the input voice.

2. The method for recognizing a state of a meridian according to claim 1 , wherein the first to fifth prompt words are respectively: “do”, “re”, “mi”, “sol” and “la”.

3. The method for recognizing a state of a meridian according to claim 1 , wherein the step of preprocessing the input voice comprises: performing noise reduction on the input voice, and performing voice signal pre-emphasis on the noise-reduced input voice.

4. The method for recognizing a state of a meridian according to claim 3 , wherein the noise reduction process comprises: removing background noise and current noise in the input voice; and the voice signal pre-emphasis comprises: improving a high frequency voice signal.

5. The method for recognizing a state of a meridian according to claim 1 , wherein the feature recognition model is established by steps of:

establishing a directed Gaussian mixture graphical model for the basic classification pitch; and

utilizing a Bayesian graphical learning algorithm to train the directed Gaussian mixture graphical model.

6. A device for recognizing a state of a meridian, wherein the state of the meridian is used to indicate a functional status of the meridian, comprising:

a receiving module, for receiving an input voice of a user;

a preprocessing module, for preprocessing the input voice;

an extraction module, for extracting a stable feature of the preprocessed input voice;

a primary classification module, for primarily classifying the stable feature on the basis of a feature recognition model, and determining a basic classification pitch, wherein the basic classification pitch comprises Gong, Shang, Jue, Zhi and Yu (respectively equivalent to do, re, mi, sol and la);

a secondary classification module, for secondarily classifying the stable feature on the basis of the feature recognition model, and determining a secondary classification tone in the basic classification pitch; and

a determination module, for recognizing the state of a meridian according to the secondary classification tone,

wherein said extracting a stable feature of the preprocessed input voice comprises:

transforming a sequence of the input voice into a frequency domain feature through Fourier transform to obtain a signal power spectrum, wherein the sequence of the input voice includes time and spatial sequences of the input voice;

utilizing discrete cosine transform to remove a correlation between signal dimensions to obtain a 24-dimensional feature of a sample, wherein the signal dimensions refer to intervals having different frequencies; and

utilizing a manifold learning method to reduce the 24-dimensional feature to be a 12-dimensional feature to obtain the stable feature,

wherein the receiving of an input voice of a user comprises:

sequentially displaying first to fifth prompt words and starting to receive first to fifth input voices of a user;

after the first to fifth input voices of the user are received, recognizing the received first to fifth input voices to be first to fifth user words;

respectively comparing the first to fifth user words with the first to fifth prompt words; and

integrating the user words matched with the prompt words to be an audio file as the input voice.

7. A mobile device, comprising a memory, a processor and a program stored in the memory and capable of running on the processor, wherein the steps of the method for recognizing the state of the meridian as claimed in claim 1 can be realized when the processor executes the program.

8. The mobile device according to claim 7 , wherein the mobile device comprises a tablet computer, a notebook, a mobile desktop computer and/or a mobile phone.