A conversion method for multiple hearing aid fitting parameters based on clinical hearing supervision

Abstract

The present invention discloses a method for converting a plurality of hearing aid fitting parameters under clinical hearing supervision, and specifically relates to the technical field of hearing aid fitting parameter conversion, the method comprising: step one, constructing user features and hearing aid features, step two, generating fitting parameters of each hearing aid at the first level, step three, generating a unified latent variable, step four, generating converted hearing aid fitting parameters, and step five, verifying the hearing aid fitting parameter conversion model. The present invention adopts user features and features of each hearing aid as input data of the fitting parameter conversion model, avoiding reliance on feedback from hearing-impaired patients during the hearing aid fitting parameter conversion process, and can output the same hearing sense fitting parameters of different hearing aids at one time according to user features and features of each hearing aid, and at the same time, fine-tunes the fitting parameter conversion model, so that the fitting parameter conversion model has self-supervised learning capability, and improves the stability of the fitting parameter conversion model.

Description

Technical Field

[0001] The present invention relates to the technical field of hearing aid fitting parameter conversion, and in particular to a method for converting a plurality of hearing aid fitting parameters under clinical hearing supervision. Background Art

[0002] Hearing aids of different appearance types and product series usually use different hearing aid chips, and different hearing aid chips also correspond to different parameter systems. There is no direct conversion tool between chip parameters. Hearing aids of different appearance types, even if they use the same hearing aid chip and are configured with the same hearing aid parameters, will have certain differences in user hearing perception due to differences in physical properties. Users will have a longer adaptation period when using hearing aids. In daily user phone replacement scenarios, hearing-impaired patients always hope to restore the hearing perception of the hearing aid before the replacement to avoid a too long adaptation process. Therefore, it is necessary to study the conversion method of various hearing aid fitting parameters.

[0003] The prior art, such as the invention patent application announced as CN103491491B, discloses a fully digital hearing aid and a non-traditional hearing aid fitting method, the method comprising: for the typical nonlinear hearing loss diagram type of the target hearing-impaired patient, at least two hearing compensation prescription modes that have been programmed with corresponding hearing aid parameters or processed with nonlinear signals are preset, and the hearing-impaired patient does not need to undergo professional fitting services of traditional digital hearing aids, including but not limited to hearing tests and hearing aid programming and debugging, but directly wears the hearing aid and selects a hearing compensation prescription mode suitable for the hearing loss type and degree of the hearing loss of the hearing-impaired patient. The volume can be adjusted in the hearing compensation prescription mode, and the hearing compensation mode and volume can be selected based on the self-perceived hearing compensation effect. The fully digital hearing aid disclosed by this invention can meet the needs of hearing-impaired patients for self-fitting and has a wide range of applications.

[0004] The prior art, such as the invention patent application CN110719558B, discloses a hearing aid fitting method, apparatus, computer equipment and storage medium, wherein the method includes: obtaining a user's voice response to an excitation sound, wherein the excitation sound is an excitation sound released in the user's ear canal; parsing the voice response through semantic analysis to obtain the response semantics; determining the user's hearing loss based on the excitation sound and the response semantics; determining the hearing compensation corresponding to the user based on the user's hearing loss; and configuring the hearing aid parameters based on the hearing compensation. The excitation sound is released, and then the user's voice response to the excitation sound is obtained, the response semantics of the voice response are parsed, the user's hearing loss is determined based on the excitation sound and the corresponding response semantics, the user's corresponding hearing compensation is determined based on the hearing loss, and the hearing aid corresponding to the user is configured based on the hearing compensation.

[0005] In combination with the above scheme, it is found that in the prior art, the conversion of hearing aid fitting parameters mostly relies on professional fitters to make adjustments based on the parameters of the hearing aid before the conversion, the hearing aid parameter indicators after the conversion and their own experience, as well as the feedback from hearing-impaired patients. On the one hand, the conversion time of the hearing aid fitting parameters is increased. On the other hand, due to the reliance on the fitter's own experience and the feedback from hearing-impaired patients, the accuracy of the hearing aid fitting parameter conversion is reduced, resulting in the hearing aid fitting parameters not being able to meet the needs of hearing-impaired patients after actual conversion. A few of them use hearing compensation and programming technology to adjust the hearing aid fitting parameters, but they still rely on the feedback from hearing-impaired patients in the process, resulting in low efficiency and accuracy in the conversion of the hearing aid fitting parameters, which makes it difficult to meet actual application needs during actual use. Summary of the invention

[0006] The purpose of the present invention is to provide a method for converting a variety of hearing aid fitting parameters under clinical hearing supervision, which solves the problems existing in the background technology.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solution: a method for converting multiple hearing aid fitting parameters based on clinical hearing supervision, the method comprising:

[0008] Step 1: Build user features and hearing aid features: Obtain the user's basic information and hearing information, build user features, obtain the information of each hearing aid, and build the features of each hearing aid.

[0009] Step 2: Generate the fitting parameters of each hearing aid at the first level: Based on user characteristics, network, generate latent variables, and input the latent variables and the characteristics of each hearing aid into In the network, the fitting parameters of each hearing aid at the first level are generated.

[0010] Step 3: Generate a unified latent variable: Match the characteristics of the hearing aid to be converted and the characteristics of the hearing aid after conversion from the characteristics of each hearing aid, and select the primary hearing aid fitting parameters corresponding to the characteristics of the hearing aid to be converted and the characteristics of the hearing aid before conversion as the conversion parameters. The input of the network generates a unified latent variable.

[0011] Step 4: Generate converted hearing aid fitting parameters: Based on the unified latent variables, the converted hearing aid features and the unified latent variables are used as The input of the network generates the fitting parameters of the converted hearing aid.

[0012] Step 5: Verify the hearing aid fitting parameter conversion model: Select the test hearing aid from the database characteristics and fitting parameters of the hearing aids The characteristics and fitting parameters of the test hearing aid are processed by the hearing aid fitting parameter conversion model to obtain the test hearing aid Fitting parameters to determine the test hearing aid and testing hearing aids Whether the fitting parameters belong to the same hearing perception fitting parameters, implement fine-tuning training.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention uses user characteristics and characteristics of each hearing aid as input data of the fitting parameter conversion model, avoiding reliance on feedback from hearing-impaired patients during the hearing aid fitting parameter conversion process, thereby improving the efficiency of hearing aid fitting parameter conversion.

[0014] 2. The present invention adopts a deep neural network algorithm, which can output the same hearing perception fitting parameters of different hearing aids at one time according to the user characteristics and the characteristics of each hearing aid, so as to facilitate the matching of the fitting parameters of the converted hearing aid during the actual hearing aid fitting parameter conversion process, thereby improving the reusability of the fitting parameter conversion model.

[0015] 3. The present invention analyzes the fitting parameters of a large number of hearing aids, the fitting parameters of hearing aids of real hearing-impaired patients, and the fitting parameters of hearing aids converted by fitters, and obtains the fitting parameters of each hearing aid with the same hearing perception as verification data for the fitting parameter conversion model, thereby improving the accuracy of the fitting parameter conversion model and meeting the needs of users in actual use.

[0016] 4. The present invention analyzes the fitting data after the conversion of the hearing aid fitting parameters with the actual hearing aid, fine-tunes the fitting parameter conversion model, refines the accuracy of the user's hearing perception replication, enables the fitting parameter conversion model to have self-supervised learning capabilities, and improves the stability of the fitting parameter conversion model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0019] Figure 2 It is a schematic diagram of the model structure of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] For ease of understanding, some professional terms involved in the present invention are explained below:

[0022] : refers to a technique commonly used in deep learning and neural networks. In a probability-based encoder-decoder architecture, The role of is to map the input to the distribution of the latent space.

[0023] : In the probability-based encoder-decoder architecture, The role of is to generate output data from samples in the latent space, usually data with a similar structure to the input data.

[0024] Fitting parameters: refers to the various audio processing parameters set according to individual hearing loss and needs during the hearing aid debugging and adjustment process. These parameters help optimize the performance of the hearing aid and ensure that the wearer gets the best hearing experience.

[0025] Hidden variables: refers to the The input data is mapped to a low-dimensional, highly abstract space, which usually represents the core features or structure of the input data rather than the original data itself. The dimension of the latent space is usually smaller than that of the original data, and the purpose is to capture the essential characteristics of the data through the learned data representation.

[0026] Spatial dimension: refers to The number of dimensions of the output vector or representation in the latent space. This dimension determines the complexity and compactness of the representation of the data after mapping from the input space to the latent space.

[0027] Same hearing perception: Hearing perception refers to the degree to which a person is accustomed to sound. Same hearing perception means that the hearing perception of the hearing aid to be converted and the hearing perception of the hearing aid after conversion should be as similar as possible to the hearing loss patient.

[0028] Back propagation algorithm: It is one of the most commonly used optimization algorithms when training neural networks. It is an optimization method based on gradient descent, which is used to calculate the gradient of the parameters (weights and biases) of each layer in the neural network, thereby optimizing the performance of the fitting parameter conversion model through gradient descent and reducing the prediction error of the fitting parameter conversion model.

[0029] Network learning rate: refers to the control of the step size of network weight adjustment at each parameter update in deep learning and neural networks. The learning rate determines how the model adjusts parameters to minimize the loss function during training.

[0030] Gradient operation: It is a partial derivative vector of a multivariable function, which represents the direction of maximum rise of the function at a certain point. In machine learning, gradient usually refers to the derivative of the loss function with respect to model parameters (such as weights and biases) calculated by the back-propagation algorithm.

[0031] Embodiment 1:

[0032] Reference Figure 1 As shown, a method for converting multiple hearing aid fitting parameters based on clinical hearing supervision, the method comprising:

[0033] Step 1: Build user features and hearing aid features: Obtain the user's basic information and hearing information, build user features, obtain the information of each hearing aid, and build the features of each hearing aid.

[0034] Step 2: Generate the fitting parameters of each hearing aid at the first level: Based on user characteristics, network, generate latent variables, and input the latent variables and the characteristics of each hearing aid into In the network, the fitting parameters of each hearing aid at the first level are generated.

[0035] Step 3: Generate a unified latent variable: Match the characteristics of the hearing aid to be converted and the characteristics of the hearing aid after conversion from the characteristics of each hearing aid, and select the primary hearing aid fitting parameters corresponding to the characteristics of the hearing aid to be converted and the characteristics of the hearing aid before conversion as the conversion parameters. The input of the network generates a unified latent variable.

[0036] Step 4: Generate converted hearing aid fitting parameters: Based on the unified latent variables, the converted hearing aid features and the unified latent variables are used as The input of the network generates the fitting parameters of the converted hearing aid.

[0037] Step 5: Verify the hearing aid fitting parameter conversion model: Select the test hearing aid from the database characteristics and fitting parameters of the hearing aids The characteristics and fitting parameters of the test hearing aid are processed by the hearing aid fitting parameter conversion model to obtain the test hearing aid Fitting parameters to determine the test hearing aid and testing hearing aids Whether the fitting parameters belong to the same hearing perception fitting parameters, implement fine-tuning training.

[0038] It should be noted that, in a specific embodiment, the database is used to store the type and chip type of each hearing aid, store the hearing sense fitting parameters of each hearing aid, and store The learning rate of the network, The initial value of .

[0039] In a specific embodiment of the present invention, the specific analysis method for constructing user characteristics is: based on the basic information pre-entered by the user, the user's age and gender characteristic data are extracted to obtain the user's basic information.

[0040] Based on the audiogram input by the user, the user's hearing threshold, type of hearing loss and level of hearing loss are extracted to obtain the user's hearing information.

[0041] It should be noted that, in a specific embodiment, the user first inputs basic information, then inputs the user's audiogram, extracts the user's hearing information, and constructs the user's features.

[0042] It should be noted that, in a specific embodiment, the influence of the user's gender on the conversion of hearing aid fitting parameters includes: men usually have more significant hearing loss in the high-frequency part, such as in the range of 2kHz-8kHz, so a higher high-frequency gain needs to be considered when converting hearing aid fitting parameters for men; women usually have obvious hearing loss in the mid-to-high frequency part, such as in the range of 1kHz-4kHz, so a more balanced gain needs to be considered when converting hearing aid fitting parameters for women to avoid excessive amplification of the high frequencies.

[0043] Older users experience high-frequency hearing loss and cannot hear high-frequency sounds clearly. Therefore, when converting hearing aid fitting parameters for older users, the gain of the high-frequency part needs to be considered. Older users have changes in sensitivity to volume. Higher volume or excessive gain may cause discomfort. When converting hearing aid fitting parameters for older users, attention needs to be paid to output limitation and volume control to ensure the comfort of the sound.

[0044] It should be noted that, in a specific embodiment, the user's audiogram is a chart used to record and evaluate a person's hearing, which is obtained through a hearing test. The horizontal axis of the audiogram represents frequency, which is in Hertz and is arranged from low frequency to high frequency. The vertical axis of the audiogram represents loudness, which is in decibels, usually ranging from 0dB to 120dB. The hearing threshold line marks the lowest volume that the tester can hear at each frequency in the audiogram. The lower the hearing threshold, the better the hearing, and the higher the hearing threshold, the more severe the hearing loss.

[0045] Types of hearing loss include conductive hearing loss, sensorineural hearing loss and mixed hearing loss. The conductive hearing loss is usually manifested on the audiogram as a higher hearing threshold in the low frequencies and normal or slight loss in the high frequencies. The sensorineural hearing loss is usually manifested on the audiogram as a higher hearing threshold in both the low and high frequencies and a smoother threshold increase. The mixed hearing loss is usually manifested on the audiogram as a conductive loss in the low frequencies and a sensorineural loss in the high frequencies.

[0046] The level of hearing loss is determined by the range of the hearing threshold. , the hearing loss is judged to be mild, and the hearing threshold is , the hearing loss is judged to be moderate, and the hearing threshold is If the hearing threshold is greater than 80, the hearing loss is considered to be severe; if the hearing threshold is greater than 80, the hearing loss is considered to be profound.

[0047] Combine the user's basic information and hearing information to construct user features, recorded as .

[0048] In a specific embodiment of the present invention, the specific analysis method for constructing the features of each hearing aid is as follows: extracting the type and chip type of each hearing aid from the database, combining the type and chip type of each hearing aid to obtain information of each hearing aid, and constructing the features of each hearing aid based on the information of each hearing aid, which is recorded as ,in , Indicates the number of each hearing aid. Indicates the total number of hearing aids.

[0049] The types of hearing aids include behind-the-ear hearing aids, in-the-ear hearing aids, in-ear hearing aids and bone conduction hearing aids.

[0050] The chip types of the hearing aids include high-end signal processing chips, mid- and low-end signal processing chips, and artificial intelligence chips.

[0051] It should be noted that in a specific embodiment, the high-end signal processing chip has the following characteristics: powerful noise management and background noise filtering functions, which can provide clear voice hearing in complex noise environments, such as Signia Xperience, which has functions such as environmental perception and voice focus, and Phonak Belong supports voice clarity and background noise management.

[0052] The mid- and low-end signal processing chips have the following characteristics: they have certain background noise filtering functions, but the effect is not as good as high-end chips. They can automatically adjust the volume in silent and noisy environments, but they usually cannot provide highly precise adjustments like high-end chips. For example, Oticon Siya provides basic sound amplification and noise filtering functions, and ReSound ENZO3D supports simple noise management and feedback suppression.

[0053] The artificial intelligence chip has the following features: it can sense the wearer's listening environment in real time and automatically adjust audio settings to optimize speech clarity and background noise suppression. Phonak Marvel AI supports intelligent speech recognition and background noise filtering. Oticon More uses deep learning technology to improve sound clarity and can adjust ambient sound perception in real time.

[0054] The present invention uses user characteristics and characteristics of each hearing aid as input data of the fitting parameter conversion model, avoids relying on feedback from hearing-impaired patients during the hearing aid fitting parameter conversion process, and improves the efficiency of hearing aid fitting parameter conversion.

[0055] In a specific embodiment of the present invention, the specific analysis method for generating hidden variables is as follows: based on Network, will The latent space dimension of the network is set to 128.

[0056] Based on user characteristics, through the model: , the user characteristics As The input data of the network passes through The network is encoded and the output latent variables are recorded as ,in express The processing of the network.

[0057] In a specific embodiment of the present invention, the specific analysis method for generating the fitting parameters of each hearing aid at the first level is: based on latent variables , characteristics of each hearing aid , through the model: , and the hidden variables With Hearing aid features As The input data of the network generates the first Hearing aid fitting parameters ,in express The network processing Indicates the input data connection symbol.

[0058] Similarly, the latent variables and the characteristics of each hearing aid are taken as The input data of the network is used to obtain the fitting parameters of each level one hearing aid.

[0059] It should be noted that, in a specific embodiment, the fitting parameters of the hearing aid include gain, maximum output sound pressure level, frequency response, feedback suppression, noise suppression, maximum gain and sensitivity.

[0060] In a specific embodiment of the present invention, the conversion The specific analysis method is as follows: Network Model: , by inputting latent variables and the first Features of hearing aids, After the decoding process of the network, we get The fitting parameters of the hearing aids will be The network is reconstructed in reverse by converting Network Model: , will Hearing aid fitting parameters and characteristics as a guide The input data of the network is used to restore the hidden variables, where represents the restored latent variable, where Indicates direction The processing of the network.

[0061] In a specific embodiment of the present invention, the specific analysis method for generating a unified latent variable is as follows: matching the features of the user's hearing aid to be converted from the features of each hearing aid, recording the features of the user's hearing aid to be converted as ,in , match the characteristics of the user's hearing aid to be transferred from the fitting parameters of each primary hearing aid to the fitting parameters of the primary hearing aid, recorded as .

[0062] By model: , generating a unified latent variable .

[0063] In a specific embodiment of the present invention, the specific analysis method for generating the fitting parameters of the converted hearing aid is as follows: matching the features of the converted hearing aid from the features of each hearing aid, recording the features of the converted hearing aid as ,in .

[0064] Based on unified latent variables , the characteristics of the hearing aid after conversion are recorded as ,pass network: , generating Hearing aid fitting parameters after hearing aid conversion .

[0065] The present invention adopts a deep neural network algorithm, which can output the same-hearing fitting parameters of different hearing aids at one time according to the user characteristics and the characteristics of each hearing aid, so as to facilitate the matching of the fitting parameters of the converted hearing aid during the actual hearing aid fitting parameter conversion process, thereby improving the reusability of the fitting parameter conversion model.

[0066] In a specific embodiment of the present invention, the determination of the test hearing aid and testing hearing aids Whether the fitting parameters belong to the same hearing sense fitting parameters, the specific analysis method is: extract the same hearing sense fitting parameters of each hearing aid from the database, recorded as ,in , , Indicates the number of the hearing aid fitting parameters. Represents the total number of hearing aid fitting parameters.

[0067] From the hearing aid fitting parameters Select a hearing aid for testing The fitting parameters are denoted as ,in , by testing hearing aids Fitting parameters The number is based on the characteristics of each hearing aid. Matching test hearing aids Corresponding features , based on the characteristics of each hearing aid Choose one of the following to test your hearing aid The characteristics of ,in .

[0068] Test-based hearing aids Features , Fitting parameters , by converting Network Model: , generating a unified latent variable .

[0069] Based on unified latent variables , Testing Hearing Aids Features ,pass Network Model: , generate the test hearing aid Fitting parameters .

[0070] Based on the hearing aid fitting parameters , Testing Hearing Aids Fitting parameters ,like , it means testing the hearing aid The test hearing aid obtained after the fitting parameters are converted Fitting parameters , and testing hearing aids Fitting parameters It belongs to the same hearing sense fitting parameters.

[0071] like , it means testing the hearing aid The test hearing aid obtained after the fitting parameters are converted Fitting parameters , and testing hearing aids Fitting parameters It does not belong to the same hearing perception fitting parameters, and performs fine-tuning training of the hearing aid fitting parameter conversion model.

[0072] The present invention analyzes the fitting parameters of a large number of hearing aids, the fitting parameters of hearing aids of real hearing-impaired patients, and the fitting parameters of hearing aids converted by fitters, and obtains the same-hearing-perception fitting parameters of each hearing aid as verification data for the fitting parameter conversion model, thereby improving the accuracy of the fitting parameter conversion model and meeting the needs of users in actual use.

[0073] In a specific embodiment of the present invention, the fine-tuning training is implemented by: It is divided into two categories, including the fitting parameters of each test hearing aid in the first category ,feature ,in , the same hearing sense fitting parameters of each test hearing aid in the second category ,feature ,in ,and , provide training data set: the fitting parameters of each test hearing aid in the first category ,feature Same hearing sense fitting parameters as the second category of each test hearing aid .

[0074] pass Network Models and Transformations Network Model: , , get the conversion fitting parameters of each test hearing aid .

[0075] Based on the same hearing perception fitting parameters of each test hearing aid in the second category Conversion fitting parameters with each tested hearing aid ,calculate and The mean square error between .

[0076] Through the back-propagation algorithm: , calculate the gradient and update Network parameters ,conduct Fine-tuning of the network, where Represents data extracted from the database The learning rate of the network, represents the gradient operation, The initial value is extracted from the database.

[0077] The present invention analyzes the fitting data after the conversion of the hearing aid fitting parameters with the actual hearing aid, fine-tunes the fitting parameter conversion model, refines the accuracy of the user's hearing perception replication, enables the fitting parameter conversion model to have self-supervised learning capabilities, and improves the stability of the fitting parameter conversion model.

[0078] Embodiment 2:

[0079] Reference Figure 2 As shown in the figure, it is a schematic diagram of the fitting parameter conversion model structure provided by the present invention. The left side is the hearing aid fitting parameter conversion part. The specific process steps are: the fitting data includes user characteristics and hearing aid characteristics, and the variational autoencoder is based on Network and Network, first realize the user characteristics through The network generates latent variables, and then the latent variables and the characteristics of each hearing aid are processed Network processing generates the fitting parameters of each hearing aid at the first level, and then selects the characteristics and fitting parameters of the hearing aid to be converted as the conversion The input end of the network generates a unified latent variable, and finally the unified latent variable and the characteristics of the transformed hearing aid are used as The input of the network generates the fitting parameters of the converted hearing aid.

[0080] Schematic diagram of the fitting parameter conversion model structure. The right side is the model verification and optimization part. Based on the same hearing perception fitting parameters, the fitting parameters and features of hearing aid a are selected as the conversion The input end of the network generates a unified latent variable, and then the unified latent variable and the features of the hearing aid b selected from the database are used as The input end of the network generates the fitting parameters of the converted hearing aid b, and determines whether the fitting parameters of the converted hearing aid b and the fitting parameters of the hearing aid a belong to the same hearing sense fitting parameters. If not, fine-tuning training is performed.

[0081] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A method for converting multiple hearing aid fitting parameters based on clinical hearing supervision, characterized in that: The method comprises: Step 1: Construct user features and hearing aid features: obtain the user's basic information and hearing information, construct user features, obtain the information of each hearing aid, and construct the features of each hearing aid; Step 2: Generate the fitting parameters of each hearing aid at the first level: Based on user characteristics, network, generate latent variables, and input the latent variables and the characteristics of each hearing aid into In the network, the fitting parameters of each hearing aid at the first level are generated; The specific analysis method for generating hidden variables is as follows: based on Network, will The latent space dimension of the network is set to 128; Based on user characteristics, through the model: , the user characteristics As The input data of the network is The network is encoded and the output latent variables are recorded as ,in express The processing of the network; The specific analysis method for generating the fitting parameters of each level one hearing aid is as follows: Based on hidden variables , characteristics of each hearing aid , through the model: , and the hidden variables With Hearing aid features As The input data of the network generates the first Hearing aid fitting parameters ,in express The network processing Indicates the input data connection symbol; Similarly, the latent variables and the characteristics of each hearing aid are taken as The input data of the network obtains the fitting parameters of each hearing aid at the first level; Step 3: Generate a unified latent variable: Match the characteristics of the hearing aid to be converted and the characteristics of the hearing aid after conversion from the characteristics of each hearing aid, and select the primary hearing aid fitting parameters corresponding to the characteristics of the hearing aid to be converted and the characteristics of the hearing aid before conversion as the conversion parameters. The input of the network generates a unified latent variable; The conversion The specific analysis method of the network is as follows: based on Network Model: , by inputting latent variables and the first Features of hearing aids, After the decoding process of the network, we get The fitting parameters of the hearing aids will be The network is reconstructed in reverse by converting Network Model: , will The fitting parameters and characteristics of each hearing aid are used as the steering The input data of the network is used to restore the hidden variables, where represents the restored latent variable, where Indicates direction The processing of the network; The specific analysis method of generating a unified hidden variable is as follows: The features of the hearing aid to be converted are matched from the features of each hearing aid, and the features of the hearing aid to be converted are recorded as ,in , match the characteristics of the user's hearing aid to be transferred from the fitting parameters of each primary hearing aid to the fitting parameters of the primary hearing aid, recorded as ; By model: , generating a unified latent variable ; Step 4: Generate converted hearing aid fitting parameters: Based on the unified latent variables, the converted hearing aid features and the unified latent variables are used as The network input generates the hearing aid fitting parameters after conversion; The specific analysis method of generating the hearing aid fitting parameters after conversion is as follows: The features of the converted hearing aids are matched from the features of each hearing aid, and the features of the converted hearing aids are recorded as ,in ; Based on unified latent variables , the characteristics of the hearing aid after conversion are recorded as ,pass network: , generating Hearing aid fitting parameters after hearing aid conversion ; Step 5: Verify the hearing aid fitting parameter conversion model: Select the test hearing aid from the database characteristics and fitting parameters of the hearing aids The characteristics and fitting parameters of the test hearing aid are processed by the hearing aid fitting parameter conversion model to obtain the test hearing aid Fitting parameters to determine the test hearing aid and testing hearing aids Whether the fitting parameters belong to the same hearing sense fitting parameters, and implement fine-tuning training; The judgment test hearing aid and testing hearing aids Whether the fitting parameters belong to the same hearing sense fitting parameters, the specific analysis method is as follows: Extract the hearing aid fitting parameters of each hearing aid from the database and record them as ,in , , Indicates the number of the hearing aid fitting parameters. Indicates the total number of hearing aid fitting parameters; From the hearing aid fitting parameters Select a hearing aid for testing The fitting parameters are recorded as ,in , by testing hearing aids Fitting parameters The number is based on the characteristics of each hearing aid. Matching test hearing aids Corresponding features , based on the characteristics of each hearing aid Choose one of the following to test your hearing aid The characteristics of ,in ; Test-based hearing aids Features , Fitting parameters , by converting Network Model: , generating a unified latent variable ; Based on unified latent variables , Testing Hearing Aids Features ,pass Network Model: , generate the test hearing aid Fitting parameters ; Based on the hearing aid fitting parameters , Testing Hearing Aids Fitting parameters ,like , it means testing the hearing aid The test hearing aid obtained after the fitting parameters are converted Fitting parameters , and testing hearing aids Fitting parameters It belongs to the same hearing sense fitting parameters; like , it means testing the hearing aid The test hearing aid obtained after the fitting parameters are converted Fitting parameters , and testing hearing aids Fitting parameters It does not belong to the same hearing perception fitting parameters, and performs fine-tuning training of the hearing aid fitting parameter conversion model.

2. The method for converting multiple hearing aid fitting parameters based on clinical hearing supervision according to claim 1, characterized in that: The specific analysis method of constructing user features is as follows: Based on the basic information pre-entered by the user, the user's age and gender are extracted to obtain the user's basic information; Based on the audiogram input by the user, extract the user's hearing threshold, the type of hearing loss and the level of hearing loss to obtain the user's hearing information; Combine the user's basic information and hearing information to construct user features, recorded as .

3. The method for converting multiple hearing aid fitting parameters based on clinical hearing supervision according to claim 2, characterized in that: The specific analysis method for constructing the characteristics of each hearing aid is as follows: The type and chip type of each hearing aid are extracted from the database, and the type and chip type of each hearing aid are combined to obtain the information of each hearing aid. Based on the information of each hearing aid, the characteristics of each hearing aid are constructed and recorded as ,in , Indicates the number of each hearing aid. Indicates the total number of hearing aids; The types of hearing aids include behind-the-ear hearing aids, in-the-ear hearing aids, in-the-ear hearing aids and bone conduction hearing aids; The chip types of the hearing aids include high-end signal processing chips, mid- and low-end signal processing chips, and artificial intelligence chips.

4. The method for converting multiple hearing aid fitting parameters based on clinical hearing supervision according to claim 1, characterized in that: The specific analysis method of implementing fine-tuning training is as follows: Fit each hearing aid to the same hearing parameters It is divided into two categories, including the fitting parameters of each test hearing aid in the first category ,feature ,in , the same hearing sense fitting parameters of each test hearing aid in the second category ,feature ,in ,and , provide training data set: the fitting parameters of each test hearing aid in the first category ,feature Same hearing sense fitting parameters as the second category of each test hearing aid ; pass Network Models and Transformations Network Model: , , get the conversion fitting parameters of each test hearing aid ; Based on the same hearing perception fitting parameters of each test hearing aid in the second category Conversion fitting parameters with each tested hearing aid ,calculate and The mean square error between ; Through the back-propagation algorithm: , calculate the gradient and update Network parameters ,conduct Fine-tuning of the network, where Represents data extracted from the database The learning rate of the network, represents the gradient operation, The initial value is extracted from the database.

Images