Method, electronic device and recording medium for compensating in-ear audio signal

1. A method for compensating an in-ear audio signal, which is applicable to an electronic device having a processor, wherein the electronic device is connected to an in-ear microphone disposed in an ear of a user and an outer microphone disposed outside the ear, the method comprising:

capturing an in-ear audio signal transmitted through an inner ear when the user speaks by using the in-ear microphone, and capturing an outer audio signal transmitted through air when the user speaks by using the outer microphone in a training stage;

establishing a machine learning model of audio signals for an objective function, and training the machine learning model by using the in-ear audio signal and the outer audio signal, comprising:

setting the in-ear audio signal as an input to an input layer of the machine learning model, and setting the outer audio signal as an input to an output layer of the machine learning model; and

training the machine learning model through the input layer and the output layer to extract an audio signal feature between the in-ear audio signal and the outer audio signal; and

converting an online stage in-ear audio signal captured by the in-ear microphone into a compensated audio signal by using the extracted audio signal feature between the in-ear audio signal and the outer audio signal in an online stage, and outputting the compensated audio signal.

2. The method for compensating an in-ear audio signal according to claim 1 , wherein the machine learning model comprises a convolutional neural network (CNN) or a deep neural network (DNN).

3. The method for compensating an in-ear audio signal according to claim 1 , wherein the step of establishing the machine learning model of audio signals for the objective function, and training the machine learning model by using the in-ear audio signal and the outer audio signal comprises:

setting the objective function based on speech quality to establish and train a speech quality model, and/or setting the objective function based on a speech recognition rate to establish and train a speech recognition model.

4. The method for compensating an in-ear audio signal according to claim 3 , further comprising:

determining to use at least one of the trained speech quality model and the trained speech recognition model to convert the in-ear audio signal into the compensated audio signal according to an application executed by the electronic device in the online stage.

5. The method for compensating an in-ear audio signal according to claim 4 , wherein if the application executed by the electronic device involves a voice call, the in-ear audio signal is converted into the compensated audio signal by using the speech quality model.

6. The method for compensating an in-ear audio signal according to claim 4 , wherein if the application executed by the electronic device involves speech recognition, the in-ear audio signal is converted into the compensated audio signal by using the speech recognition model.

7. The method for compensating an in-ear audio signal according to claim 6 , further comprising:

performing speech recognition on the converted compensated audio signal, and outputting a speech recognition result.

8. An electronic device comprising:

a connection device, connected to an in-ear microphone disposed in an ear of a user and an outer microphone disposed outside the ear;

a storage device, storing one or more commands; and

a processor, coupled to the connection device and the storage device, configured to execute the commands to:

capture an in-ear audio signal transmitted through an inner ear when the user speaks by using the in-ear microphone, and capture an outer audio signal transmitted through air when the user speaks by using the outer microphone in a training stage;

establish a machine learning model of audio signals for an objective function, and train the machine learning model by using the in-ear audio signal and the outer audio signal, wherein the processor is configured to set the in-ear audio signal as an input to an input layer of the machine learning model, set the outer audio signal as an input to an output layer of the machine learning model, and train the machine learning model through the input layer and the output layer to extract an audio signal feature between the in-ear audio signal and the outer audio signal; and

convert an online stage in-ear audio signal captured by the in-ear microphone into a compensated audio signal by using the extracted audio signal feature between the in-ear audio signal and the outer audio signal in an online stage, and output the compensated audio signal.

9. The electronic device according to claim 8 , wherein the machine learning model comprises a convolutional neural network or a deep neural network.

10. The electronic device according to claim 8 , wherein the processor is configured to set the objective function based on speech quality to establish and train a speech quality model, and/or set the objective function based on a speech recognition rate to establish and train a speech recognition model.

11. The electronic device according to claim 10 , wherein the processor is further configured to execute an application.

12. The electronic device according to claim 11 , wherein the processor is further configured to determine to use at least one of the trained speech quality model and the trained speech recognition model to convert the in-ear audio signal into the compensated audio signal according to the application executed by the processor in the online stage.

13. The electronic device according to claim 12 , wherein if the application executed by the processor involves a voice call, the in-ear audio signal is converted into the compensated audio signal by using the speech quality model.

14. The electronic device according to claim 12 , wherein if the application executed by the processor involves speech recognition, the in-ear audio signal is converted into the compensated audio signal by using the speech recognition model.

15. The electronic device according to claim 14 , wherein the processor is further configured to perform speech recognition on the converted compensated audio signal, and output a speech recognition result.

16. A computer-readable recording medium, recording a program loaded into a processor to perform steps below:

capturing an in-ear audio signal transmitted through an inner ear when a user speaks by using an in-ear microphone, and capturing an outer audio signal transmitted through air when the user speaks by using an outer microphone in a training stage;

establishing a machine learning model of audio signals for an objective function, and training the machine learning model by using the in-ear audio signal and the outer audio signal, comprising:

setting the in-ear audio signal as an input to an input layer of the machine learning model, and setting the outer audio signal as an input to an output layer of the machine learning model; and

training the machine learning model through the input layer and the output layer to extract an audio signal feature between the in-ear audio signal and the outer audio signal; and

converting an online stage in-ear audio signal captured by the in-ear microphone into a compensated audio signal by using the extracted audio signal feature between the in-ear audio signal and the outer audio signal in an online stage, and outputting the compensated audio signal.