Hearing aid noise reduction method and device

Abstract

The embodiment of the invention relates to the technical field of hearing aids, and discloses a hearing aid noise reduction method and device, and the method comprises the steps: collecting environment audio data through a hearing aid; inputting the environment audio data into a preset audio recognition model, and obtaining target noise data through the audio recognition model; and performing visual adjustment on target noise data in the environment audio data to generate target audio data. According to the method, the noise source can be actively identified, the frequency range and the loudness of the noise source are obtained, the noise source is selected visually, and the loudness of the noise source in the hearing aid is adjusted in real time, so that discomfort of hearing aid caused by continuous output of periodic environmental noise to human ears is avoided, and the hearing aid is more comfortable to use. And the condition of tinnitus is avoided. Meanwhile, through visual operation of the hearing aid, the user can operate the hearing aid by himself / herself very conveniently, and the applicability of the hearing aid is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of hearing aids, and more particularly to a hearing aid noise reduction method and device. Background Art

[0002] A hearing aid is a small loudspeaker that amplifies sounds that are not audible, and then uses the residual hearing of the hearing-impaired to send the sounds to the auditory center of the brain, so that the hearing-impaired can feel the sounds. This brings great convenience to the hearing-impaired. A hearing aid is mainly composed of five parts: microphone, amplifier, earphone, power supply and volume control. Hearing aids are divided into air conduction hearing aids and bone conduction hearing aids according to the conduction method; and are classified into box type, glasses type, hairpin type, behind-the-ear type, in-the-ear type, ear canal type, and deep ear canal type hearing aids according to the usage method.

[0003] In the prior art, periodic environmental noise in hearing aids, such as the periodic working sound of air-conditioning motors, external periodic insect sounds, etc., can easily amplify external sounds to the hearing aid users, causing great discomfort to the hearing aid users when obtaining external audio, and in severe cases, it can easily cause tinnitus in the users. Summary of the invention

[0004] In view of the above problems, an embodiment of the present invention provides a hearing aid noise reduction method and device, which is used to solve the problem of periodic environmental noise in hearing aids in the prior art, such as the periodic working sound of air-conditioning motors, external periodic insect sounds, etc., which easily amplify external sounds to the hearing aid users, causing great discomfort to the hearing aid users when obtaining external audio, and in serious cases, easily causing tinnitus in the users.

[0005] According to one aspect of an embodiment of the present invention, a hearing aid noise reduction method is provided, the method comprising: Collecting ambient audio data through hearing aids; Inputting the environmental audio data into a preset audio recognition model, and obtaining target noise data through the audio recognition model; Visually adjust the target noise data in the environmental audio data to generate target audio data.

[0006] In some optional ways, inputting the environmental audio data into a preset audio recognition model, and obtaining the target noise data through the audio recognition model includes: The environmental audio input is used as the input of the audio recognition model, and the target noise data is output after being processed by the audio recognition model; wherein the target noise data includes the type of noise source, the frequency of the noise source, and the loudness of the noise source.

[0007] In some optional embodiments, the audio recognition model is trained by: Collect data from multiple noise sources; Extracting features from the noise source data using an MFCCs feature extraction algorithm to obtain MFCC parameters; The MFCC parameters are trained and classified through the HMM model.

[0008] In some optional embodiments, extracting features from the noise source data using an MFCCs feature extraction algorithm to obtain MFCC parameters specifically includes: The noise source data is framed and processed using Hamming windowing; The frequency spectrum of the noise source data is obtained by Fourier transform, and the speech energy spectrum in the frequency spectrum is processed by a Mel frequency filter group, and then subjected to inverse Fourier transform to obtain MFCC parameters.

[0009] In some optional embodiments, the training and classification of the MFCC parameters by the HMM model specifically includes: According to the type of noise source data, set the corresponding number of HMM models and define the forward probability variable of each HMM model as α t (i), in the HMM model parameter λ i Under the condition, at the tth moment, it is in state S i , output the first part of the observation sequence O = O1, O2, ..., O t The output probability is: α t (i) = P{O1, O2, ..., O t ,q t =S i |λ}(1≤t≤T,1≤i≤N) Among them, the maximum value of the output probability is the type of noise source data corresponding to the HMM model.

[0010] In some optional embodiments, the forward probability variable α t (i) and the forward algorithm for output probability P(O|λ) includes the following steps: S1, for 1≤i≤N, the initialization α of the joint probability under the condition of being in state Si and the initial observation variable being O1 t (i) = π i b i (O1); S2, for 1≤t≤T-1,1≤j≤N, there is a t+1 (j)=[ ] j (O t+1 ), the iteration relationship at time t is: the state S at time t i and the state S at time t+1j The conversion relationship between them; S3, through the recursion of step S2, obtain the forward probability variable value α t (i), then according to α t (i) Calculate the final output probability P(O|λ) = .

[0011] In some optional methods, after the MFCC parameters are trained and classified through the HMM model, the maximum output probability and the second maximum output probability are obtained, and the maximum output probability and the second maximum output probability are output to the SVM model for a second training and classification to obtain the final target noise data.

[0012] In some optional manners, visually adjusting the target noise data in the environmental audio data specifically includes: The environmental audio data is graphically represented in the audio analysis graph, and the frequency band and loudness of the target noise data are marked in the audio analysis graph of the environmental audio data; The target noise data is framed and the loudness of the target noise data in the hearing aid is set to generate target audio data.

[0013] In some optional embodiments, collecting the environmental audio data through the hearing aid specifically includes: acquiring the environmental audio data through the microphone of the hearing aid within a preset range of the noise source.

[0014] According to another aspect of an embodiment of the present invention, a hearing aid noise reduction device is provided, comprising: An audio acquisition module, used to collect environmental audio data through a hearing aid; A noise recognition module, used to input the environmental audio data into a preset audio recognition model, and obtain target noise data through the audio recognition model; and a noise reduction module, which is used to visually adjust the target noise data in the environmental audio data to generate target audio data.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the embodiment of the present invention collects environmental audio data through a hearing aid; inputs the environmental audio data into a preset audio recognition model, and obtains target noise data through the audio recognition model; and visually adjusts the target noise data in the environmental audio data to generate target audio data. This method can be used to actively identify the noise source, obtain the frequency range and loudness of the noise source, and visually select the noise source and adjust the loudness of the noise source in the hearing aid in real time to avoid periodic environmental noise from being continuously output to the human ear, causing hearing discomfort and tinnitus. At the same time, through the visual operation of the hearing aid, it is greatly convenient for users to operate by themselves, thereby improving the applicability of the hearing aid.

[0016] The above description is only an overview of the technical solution of the embodiment of the present invention. In order to more clearly understand the technical means of the embodiment of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present invention. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings: Figure 1 A schematic diagram showing the flow of the hearing aid adjustment method based on otoacoustic emissions provided by the present invention is shown; Figure 2 A schematic diagram of a process of performing otoacoustic emission testing on a target user provided by the present invention is shown; Figure 3 The schematic diagram of the structure of the hearing aid adjustment device based on otoacoustic emissions provided by the present invention is shown.

[0018] Figure 4 The schematic diagram of the structure of the testing device provided by the present invention is shown. DETAILED DESCRIPTION

[0019] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0020] Embodiment 1, Figure 1 The first embodiment of a hearing aid noise reduction method of the present invention is shown. The present application is based on a hearing aid and is applied in the adjustment of a hearing aid in the presence of periodic noise. The method specifically includes: 110, collecting environmental audio data through a hearing aid; In step 110, collecting environmental audio data through a hearing aid specifically includes: acquiring environmental audio data through a microphone of the hearing aid within a preset range of the noise source.

[0021] 120, input the environmental audio data into a preset audio recognition model, and obtain target noise data through the audio recognition model; in step 120, input the environmental audio data into a preset audio recognition model, and obtain target noise data through the audio recognition model, including: using the environmental audio input as the input of the audio recognition model, and outputting the target noise data after processing through the audio recognition model; wherein the target noise data includes the type of noise source, the frequency of the noise source, and the loudness of the noise source.

[0022] 130, visually adjust the target noise data in the ambient audio data to generate the target audio data. In step 130, the target noise data in the ambient audio data is visually adjusted, specifically including: graphically illustrating the ambient audio data in the audio analysis graph, and marking the frequency band and loudness of the target noise data in the audio analysis graph of the ambient audio data; box-selecting the target noise data, and setting the loudness of the target noise data in the hearing aid to generate the target audio data.

[0023] In this embodiment, the embodiment of the present invention collects environmental audio data through a hearing aid; inputs the environmental audio data into a preset audio recognition model, and obtains target noise data through the audio recognition model; and visually adjusts the target noise data in the environmental audio data to generate target audio data. This method can be used to actively identify the noise source, obtain the frequency range and loudness of the noise source, and visually select the noise source and adjust the loudness of the noise source in the hearing aid in real time to avoid periodic environmental noise from being continuously output to the human ear, causing hearing discomfort and tinnitus. At the same time, through the visual operation of the hearing aid, it is greatly convenient for users to operate by themselves, thereby improving the applicability of the hearing aid.

[0024] Embodiment 2: Based on Example 1, the present application provides a second embodiment of a hearing aid noise reduction method. This embodiment further describes Example 1.

[0025] In some optional methods, the environmental audio data is input into a preset audio recognition model, and the target noise data is obtained through the audio recognition model, including: using the environmental audio input as the input of the audio recognition model, processing it through the audio recognition model and outputting the target noise data; wherein the target noise data includes the type of noise source, the frequency of the noise source, and the loudness of the noise source.

[0026] The audio recognition model is trained in the following way: 210, collecting multiple noise source data; in step 210, the noise source data can be divided into training data and test data, the model is trained by the training data, and the model is verified by the test data.

[0027] 220, feature extraction is performed on the noise source data through the MFCCs feature extraction algorithm to obtain MFCC parameters; in step 220, feature extraction is performed on the noise source data through the MFCCs feature extraction algorithm to obtain MFCC parameters, specifically including: framing the noise source data and using Hamming windowing processing; obtaining the spectrum of the noise source data through Fourier transform, and processing the speech energy spectrum in the spectrum through the Mel frequency filter group, and then obtaining MFCC parameters through inverse Fourier transform.

[0028] 230, training and classifying the MFCC parameters through the HMM model. In step 230, training and classifying the MFCC parameters through the HMM model specifically includes: setting a corresponding number of HMM models according to the type of noise source data, and defining the forward probability variable of each HMM model as α t (i), in the HMM model parameter λ i Under the condition, at the tth moment, it is in state S i , output the first part of the observation sequence O = O1, O2, ..., O t Output probability: α t (i) = P{O1, O2, ..., O t ,q t =S i |λ}(1≤t≤T,1≤i≤N) where the maximum value of the output probability is the type of noise source data corresponding to the HMM model.

[0029] More specifically, in step 220, the noise source data is feature extracted by the MFCCs feature extraction algorithm to obtain MFCC parameters, which specifically includes: framing the noise source data and using Hamming windowing processing; obtaining the spectrum of the noise source data by Fourier transform, and processing the speech energy spectrum in the spectrum by Mel frequency filter group, and then obtaining MFCC parameters by inverse Fourier transform. Among them, framing is to divide the speech signal into multiple short time frames, usually 20-40 milliseconds per frame, and there is 50% overlap between frames. This is done to assume that the speech signal is short-term stable in a short time, so that each frame is processed independently. Using Hamming windowing processing is to apply a Hamming window to each frame, increase the continuity of the left and right ends of the frame, and reduce the edge effect. Fourier transform the noise source data and perform FFT on each frame signal to obtain a frequency domain representation. Processing the speech energy spectrum in the spectrum by Mel frequency filter group specifically includes applying Mel filter group, the frequency response of these filters imitates the nonlinear characteristics of the human auditory system. Each filter calculates the energy of its corresponding frequency band. The inverse Fourier transform calculates the corresponding original function f(t) for a given Fourier transform F(jω). It is the inverse process of the Fourier transform and is used to convert the frequency domain signal back to the time domain signal to obtain the MFCC parameters.

[0030] It should be noted that no preprocessing is required during the MFCCs feature extraction process to avoid loss of periodic noise data.

[0031] More specifically, in step 230, the MFCC parameters are trained and classified by the HMM model, specifically including: According to the type of noise source data, set the corresponding number of HMM models and define the forward probability variable of each HMM model as α t (i), in the HMM model parameter λ i Under the condition, at the tth moment, it is in state S i , output the first part of the observation sequence O = O1, O2, ..., O t The output probability is: α t (i) = P{O1, O2, ..., O t ,q t =S i |λ}(1≤t≤T,1≤i≤N) Among them, the maximum value of the output probability is the type of noise source data corresponding to the HMM model.

[0032] In the above steps, the forward probability variable α t (i) and the forward algorithm for output probability P(O|λ) includes the following steps: S1, for 1≤i≤N, the initialization α of the joint probability under the condition of being in state Si and the initial observation variable being O1 t (i) = π i b i (O1); S2, for 1≤t≤T-1,1≤j≤N, there is a t+1 (j)=[ ] j (O t+1 ), the iteration relationship at time t is: the state S at time t i and the state S at time t+1 j The conversion relationship between them; S3, through the recursion of step S2, obtain the forward probability variable value α t (i), then according to α t (i) Calculate the final output probability P(O|λ) = .

[0033] Through this application, different output probability values ​​of different HMM models can be obtained, and finally the value with the largest output probability is selected as the target noise data.

[0034] In addition, after the MFCC parameters are trained and classified by the HMM model, the maximum output probability and the second maximum output probability are obtained, and the maximum output probability and the second maximum output probability are output to the SVM model for a second training and classification to obtain the final target noise data. In this embodiment, the HMM model can also be judged and classified twice by the SVM model to further improve the accuracy of classification. The results of the HMM model are optimized by the SVM model, and the overall noise source recognition accuracy reaches more than 96%.

[0035] In some optional manners, visually adjusting the target noise data in the environmental audio data specifically includes: 310, graphically displaying the environmental audio data in an audio analysis graph, and marking the frequency band and loudness of the target noise data in the audio analysis graph of the environmental audio data; 320 , select the target noise data, set the loudness of the target noise data in the hearing aid, and generate target audio data.

[0036] In this embodiment, the hearing aid noise reduction method of the present application can be applied to a noise reduction device, wherein the noise reduction device may include a hearing aid and a test terminal connected to the hearing aid, wherein the test terminal may be a mobile phone, a tablet, a computer host, etc., the hearing aid is provided with a first communication module, and the test terminal is provided with a second communication module, the first communication module is connected to the second communication module, so that the environmental audio data collected by the hearing aid is transmitted to the test terminal, and the test terminal first outputs the environmental audio data to the audio recognition model to identify the target noise data. When the environmental audio data is subjected to audio analysis through the test terminal, the target noise data can be marked, and the marked target noise data can be selected and its loudness can be adjusted to achieve visual adjustment of the noise source, and finally generate the environmental frequency data after noise reduction. In addition, this embodiment can specifically select the noise source, and then adjust the amplification degree of the noise source in the hearing aid to achieve overall noise reduction of the noise source.

[0037] Embodiment 3: According to another aspect of an embodiment of the present invention, a hearing aid noise reduction device 400 is provided, comprising: The audio collection module 410 is used to execute step 110 of embodiment 1; specifically, it includes: collecting environmental audio data through the hearing aid specifically includes: obtaining environmental audio data through the microphone of the hearing aid within a preset range of the noise source.

[0038] The noise recognition module 420 is used to execute step 120 of Example 1; specifically including: inputting the environmental audio data into a preset audio recognition model, and obtaining target noise data through the audio recognition model, including: using the environmental audio input as the input of the audio recognition model, and outputting the target noise data after processing through the audio recognition model; wherein the target noise data includes the type of noise source, the frequency of the noise source, and the loudness of the noise source.

[0039] and a noise reduction module 430, for executing step 130 of embodiment 1. Specifically comprising: visually adjusting the target noise data in the environmental audio data, specifically comprising: graphically illustrating the environmental audio data in the audio analysis graph, and marking the frequency band and loudness of the target noise data in the audio analysis graph of the environmental audio data; selecting the target noise data, and setting the loudness of the target noise data in the hearing aid to generate the target audio data.

[0040] In this embodiment, the embodiment of the present invention collects environmental audio data through a hearing aid; inputs the environmental audio data into a preset audio recognition model, and obtains target noise data through the audio recognition model; and visually adjusts the target noise data in the environmental audio data to generate target audio data. This method can be used to actively identify the noise source, obtain the frequency range and loudness of the noise source, and visually select the noise source and adjust the loudness of the noise source in the hearing aid in real time to avoid periodic environmental noise from being continuously output to the human ear, causing hearing discomfort and tinnitus. At the same time, through the visual operation of the hearing aid, it is greatly convenient for users to operate by themselves, thereby improving the applicability of the hearing aid.

[0041] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system or other device. In addition, the embodiments of the present invention are not directed to any particular programming language.

[0042] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. Similarly, in order to simplify the present invention and help understand one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. Wherein, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself is a separate embodiment of the present invention.

[0043] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive.

[0044] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.

Claims

1. A hearing aid noise reduction method, characterized in that: The method comprises: Collecting ambient audio data through hearing aids; Inputting the environmental audio data into a preset audio recognition model, and obtaining target noise data through the audio recognition model; Visually adjust the target noise data in the environmental audio data to generate target audio data.

2. The hearing aid noise reduction method according to claim 1, characterized in that: Inputting the environmental audio data into a preset audio recognition model, and obtaining target noise data through the audio recognition model, including: The environmental audio input is used as the input of the audio recognition model, and the target noise data is output after being processed by the audio recognition model; wherein the target noise data includes the type of noise source, the frequency of the noise source, and the loudness of the noise source.

3. The hearing aid noise reduction method according to claim 2, characterized in that: The audio recognition model is trained in the following way: Collect data from multiple noise sources; Extracting features from the noise source data using an MFCCs feature extraction algorithm to obtain MFCC parameters; The MFCC parameters are trained and classified through the HMM model.

4. The hearing aid noise reduction method according to claim 3, characterized in that: The feature extraction of the noise source data by using the MFCCs feature extraction algorithm to obtain MFCC parameters specifically includes: The noise source data is framed and processed using Hamming windowing; The frequency spectrum of the noise source data is obtained by Fourier transform, and the speech energy spectrum in the frequency spectrum is processed by a Mel frequency filter group, and then subjected to inverse Fourier transform to obtain MFCC parameters.

5. The hearing aid noise reduction method according to claim 4, characterized in that: The training and classification of the MFCC parameters by the HMM model specifically includes: According to the type of noise source data, set the corresponding number of HMM models and define the forward probability variable of each HMM model as α t (i), in the HMM model parameter λ i Under the condition, at the tth moment, it is in state S i , output the first part of the observation sequence O = O1, O2, ..., O t The output probability is: α t (i)=P{O1,O2,…,O t ,q t =S i |λ}(1≤t≤T,1≤i≤N) Among them, the maximum value of the output probability is the type of noise source data corresponding to the HMM model.

6. The hearing aid noise reduction method according to claim 5, characterized in that: The forward probability variable α t (i) and the forward algorithm for output probability P(O|λ) consists of the following steps: S1, for 1≤i≤N, the initialization α of the joint probability under the condition of being in state Si and the initial observation variable being O1 t (i) = π i b i (O1); S2, for 1≤t≤T-1,1≤j≤N, there is a t+1 (j)=[ ] j (O t+1 ), the iteration relationship at time t is: the state S at time t i and the state S at time t+1 j The conversion relationship between them; S3, through the recursion of step S2, obtain the forward probability variable value α t (i), then according to α t (i) Calculate the final output probability P(O|λ) = .

7. The hearing aid noise reduction method according to claim 6, characterized in that: After the MFCC parameters are trained and classified by the HMM model, the maximum output probability and the second maximum output probability are obtained, and the maximum output probability and the second maximum output probability are output to the SVM model for a second training and classification to obtain the final target noise data.

8. The hearing aid noise reduction method according to claim 2, characterized in that: Visually adjusting the target noise data in the environmental audio data includes: The environmental audio data is graphically represented in the audio analysis graph, and the frequency band and loudness of the target noise data are marked in the audio analysis graph of the environmental audio data; The target noise data is framed and the loudness of the target noise data in the hearing aid is set to generate target audio data.

9. The hearing aid noise reduction method according to claim 1, characterized in that: The collecting of environmental audio data through the hearing aid specifically includes: acquiring the environmental audio data through the microphone of the hearing aid within a preset range of the noise source.

10. A hearing aid noise reduction device, characterized in that: include: An audio acquisition module, used to collect environmental audio data through a hearing aid; A noise recognition module, used to input the environmental audio data into a preset audio recognition model, and obtain target noise data through the audio recognition model; and a noise reduction module, which is used to visually adjust the target noise data in the environmental audio data to generate target audio data.

Images