Hearing assessment method and device based on fixed stimulation intensity, terminal and medium

Abstract

The invention relates to the technical field of hearing testing, in particular to a hearing evaluation method and device based on fixed stimulation intensity, a terminal and a medium, and the method comprises the steps: obtaining auditory brainstem response waves collected under preset fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves; extracting waveform features of the auditory brainstem reaction waves, converting the waveform features into waveform feature vectors, and converting the demographic features into demographic feature vectors; and inputting the waveform feature vector and the demographic feature vector into a trained auditory evaluation model, and outputting an auditory evaluation result which comprises any one or more of a hearing loss level classification result, a hearing loss property classification result and a hearing graph classification result. According to the method and the device, the test times required for evaluating the auditory level are reduced, the auditory evaluation time is shortened, and the recognition result of one or more dimensions is output based on the auditory evaluation model, so that the auditory evaluation accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of hearing testing, and in particular to a hearing assessment method, device, terminal and medium based on fixed stimulation intensity. Background Art

[0002] The current main auditory assessment technologies, such as pure tone audiometry and speech testing, require the subjective cooperation of the subjects during the test process, and their application is subject to certain limitations.

[0003] The Auditory Brainstem Response (ABR) is a seven-wave response evoked by acoustic stimulation and recorded from the scalp surface with a latency of 1 to 10 milliseconds. Because the threshold of the ABR is significantly correlated with that of pure-tone audiometry, and because the test results are less susceptible to factors such as the subject's subjective awareness, emotions, educational level, responsiveness, and behavioral coordination, the ABR test has been widely used in auditory assessment.

[0004] However, since auditory assessment based on auditory brainstem response requires collecting response waves under different stimulus intensities and performing auditory assessment based on response waves under different stimulus intensities, auditory assessment based on auditory brainstem response is time-consuming, and the waveform differentiation at the threshold level is poor, and recognition is more dependent on the tester's experience. Therefore, the efficiency and accuracy of auditory assessment are low.

[0005] Therefore, the existing technology has defects and needs to be improved and developed. Summary of the Invention

[0006] The present application provides a method, device, terminal and medium for auditory assessment based on fixed stimulation intensity to solve the technical problem of low efficiency and accuracy of auditory assessment in related technologies.

[0007] To achieve the above objectives, this application adopts the following technical solutions:

[0008] A method for auditory assessment based on fixed stimulus intensity, wherein the method comprises:

[0009] obtaining auditory brainstem response waves collected at a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves;

[0010] extracting waveform features of the auditory brainstem response wave, converting the waveform features into waveform feature vectors, and converting the demographic features into demographic feature vectors;

[0011] The waveform feature vector and the demographic feature vector are input into a trained hearing assessment model to output a hearing assessment result, wherein the hearing assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result.

[0012] In one embodiment of the present application, the predetermined fixed stimulation intensity is 80 dB nHL.

[0013] In one embodiment of the present application, the waveform characteristics include: one or more of: the latency of the predetermined reaction wave, the amplitude corresponding to the predetermined reaction wave, the interwave interval between the predetermined reaction waves, and the amplitude ratio between the predetermined reaction waves; the demographic characteristics include: the gender and / or age of the subject.

[0014] In one embodiment of the present application, the training step of the auditory assessment model includes:

[0015] Obtaining a pre-constructed training set, the training set comprising a plurality of auditory brainstem response wave training data collected under a predetermined fixed stimulation intensity, and demographic characteristic training data, hearing loss level classification labels, hearing loss nature classification labels, and hearing pattern classification labels corresponding to each auditory brainstem response wave training data;

[0016] Extracting waveform feature training data corresponding to the auditory brainstem response wave training data, converting the waveform feature training data into waveform feature vector training data, and converting the demographic feature training data into demographic feature vector training data;

[0017] training a pre-built hearing loss degree classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss level classification labels; training a pre-built hearing loss property classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss property classification labels; and training a hearing pattern classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing pattern classification labels;

[0018] The trained hearing loss degree classification model, hearing loss nature classification model and hearing graph classification model are integrated into an auditory assessment model.

[0019] In one embodiment of the present application, the hearing loss degree classification model, the hearing loss nature classification model, and the hearing pattern classification model are all multi-layer feedforward neural networks.

[0020] In one embodiment of the present application, the hearing loss degree classification model and the hearing loss nature classification model both adopt a four-layer neural network architecture, and the dimension of the hidden layer of the hearing loss degree classification model and the hearing loss nature classification model is 128 dimensions; the hearing graph classification model adopts a five-layer neural network architecture, and the dimension of the hidden layer of the hearing graph classification model is 256 dimensions.

[0021] In one embodiment of the present application, during the training process of the hearing loss degree classification model, the hearing loss nature classification model, and the hearing pattern classification model, a cross entropy loss function is used for classification tasks, and an adaptive moment estimation optimizer is used to update the model parameters.

[0022] The present application also provides a hearing assessment device based on fixed stimulation intensity, wherein the device comprises:

[0023] an acquisition module, configured to acquire auditory brainstem response waves collected under a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves;

[0024] a conversion module, configured to extract waveform features of the auditory brainstem response wave, convert the waveform features into waveform feature vectors, and convert the demographic features into demographic feature vectors;

[0025] An output module is used to input the waveform feature vector and the demographic feature vector into a trained hearing assessment model and output a hearing assessment result, wherein the hearing assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result.

[0026] The present application also provides a terminal, which includes: a memory, a processor, and an auditory assessment program based on fixed stimulation intensity stored in the memory and executable on the processor. When the auditory assessment program based on fixed stimulation intensity is executed by the processor, the steps of the auditory assessment method based on fixed stimulation intensity as described above are implemented.

[0027] The present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program can be executed to implement the steps of the auditory assessment method based on fixed stimulation intensity as described above.

[0028] Beneficial effects of the present invention: The method of the embodiment of the present invention obtains the auditory brainstem response wave collected at a predetermined fixed stimulation intensity and the demographic characteristics corresponding to the auditory brainstem response wave; extracts the waveform characteristics of the auditory brainstem response wave, converts the waveform characteristics into a waveform feature vector, and converts the demographic characteristics into a demographic feature vector; inputs the waveform feature vector and the demographic feature vector into a trained auditory assessment model, and outputs an auditory assessment result, wherein the auditory assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result. The present application reduces the number of tests required to assess the auditory level by processing the auditory brainstem response data collected at a predetermined fixed stimulation intensity, thereby shortening the auditory assessment time. Moreover, the accuracy of the auditory assessment is improved by outputting the recognition results of one or more dimensions of the hearing loss level classification result, the hearing loss nature classification result, and the hearing pattern classification result based on the auditory assessment model. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of a preferred embodiment of the auditory assessment method based on fixed stimulation intensity in the present invention.

[0030] Figure 2 It is the accuracy comparison result before and after feature screening in the present invention.

[0031] Figure 3 This is a feature ablation experiment result diagram after feature screening in the present invention.

[0032] Figure 4 It is the auditory brainstem response wave collected at the 80dB nHL stimulation sound intensity in the present invention.

[0033] Figure 5 It is a schematic diagram of characteristic analysis of auditory brainstem response waves collected at 80dB nHL stimulus intensity in the present invention.

[0034] Figure 6 This is the waveform characteristic of the auditory brainstem response wave in the present invention.

[0035] Figure 7 It is a characteristic graph obtained after data cleaning and standardization of the waveform characteristic vector and the demographic characteristic vector in the present invention.

[0036] Figure 8 It is a schematic diagram of the neural network architecture in the present invention.

[0037] Figure 9 This is a functional principle block diagram of a preferred embodiment of the auditory assessment device based on fixed stimulation intensity in the present invention.

[0038] Figure 10It is a functional principle block diagram of a preferred embodiment of the terminal in the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0040] The following describes the auditory assessment method, device, terminal and medium based on fixed stimulation intensity according to an embodiment of the present application with reference to the accompanying drawings. In response to the problem of low efficiency and accuracy of auditory assessment in the related art mentioned in the above background technology, the present application provides an auditory assessment method based on fixed stimulation intensity, in which the auditory brainstem response wave collected at a predetermined fixed stimulation intensity and the demographic characteristics corresponding to the auditory brainstem response wave are obtained; the waveform characteristics of the auditory brainstem response wave are extracted, the waveform characteristics are converted into a waveform feature vector, and the demographic characteristics are converted into a demographic feature vector; the waveform feature vector and the demographic feature vector are input into a trained auditory assessment model, and the auditory assessment result is output, wherein the auditory assessment result includes any one or more of the hearing loss level classification result, the hearing loss nature classification result and the hearing pattern classification result. This application reduces the amount of response wave data by processing auditory brainstem response data collected under a predetermined fixed stimulation intensity, thereby shortening the auditory assessment time. In addition, based on the auditory assessment model, the application outputs the recognition results of one or more dimensions of the hearing loss level, hearing loss nature, and hearing graph, thereby improving the accuracy of the auditory assessment.

[0041] See Figure 1 The auditory assessment method based on fixed stimulation intensity according to an embodiment of the present invention comprises the following steps:

[0042] Step S100: Acquire auditory brainstem response waves collected under a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves.

[0043] In an embodiment of the present application, the predetermined fixed stimulation intensity is 80dB nHL. In the existing technology, the response waves under different stimulation intensities are generally obtained, and the auditory assessment is performed based on the response waves under different stimulation intensities, which results in a longer time-consuming auditory assessment and poor waveform differentiation at the threshold level. The present application found that under normal circumstances, the auditory brainstem response (ABR) waveform induced by the 80dB nHL stimulation sound intensity is well differentiated and easier to clinically identify. Therefore, if the hearing loss of the subject can be assessed by the auditory brainstem response data under a fixed stimulation intensity of 80dBnHL, the time for auditory assessment can be effectively shortened to meet the needs of auditory assessment for a large population, thereby improving the efficiency of auditory assessment.

[0044] like Figure 1 As shown, the auditory assessment method based on fixed stimulation intensity further includes the following steps:

[0045] Step S200: extracting waveform features of the auditory brainstem response wave, converting the waveform features into waveform feature vectors, and converting the demographic features into demographic feature vectors.

[0046] In an embodiment of the present application, the waveform characteristics include: one or more of: the latency of the predetermined reaction wave, the amplitude corresponding to the predetermined reaction wave, the interwave interval between the predetermined reaction waves, and the amplitude ratio between the predetermined reaction waves; the demographic characteristics include: the gender and / or age of the subject.

[0047] Specifically, the auditory brainstem response waves usually include: wave I, wave II, wave III, wave IV, wave V, wave VI, and wave VII, a total of 7 waves. The predetermined response waves of this application refer to: wave I, wave III, and wave V. The source of wave I is near the end of the auditory nerve, with a high occurrence rate. It is the main parameter wave for analyzing the auditory brainstem response; the source of wave III is the ipsilateral cochlear nucleus, and a small part of the eighth cranial nerve fibers are involved. It has a high occurrence rate and is the main parameter wave for analyzing the auditory brainstem response; the source of wave V is the positive component originating from the lateral lemniscus and the negative component originating from the inferior colliculus. It is the most stable and usually the wave with the highest amplitude.

[0048] Therefore, the latency of the predetermined reaction waves includes: wave I latency, wave III latency, and wave V latency; the amplitudes corresponding to each predetermined reaction wave include: wave I amplitude, wave III amplitude, and wave V amplitude; the interwave intervals between each reaction wave include: wave I-III interval, wave III-V interval, and wave IV interval; the amplitude ratios between each reaction wave include: wave I-III amplitude ratio, wave III-V amplitude ratio, and wave IV amplitude ratio.

[0049] According to the results of the correlation analysis, the waveform features of this application selected the III-V wave interval, the V wave latency and the IV wave interval, and the demographic feature selected the age of the subject. Compared with using all waveform features, the accuracy of auditory assessment can be improved. The accuracy comparison results before and after feature screening are shown in the figure. Figure 2 The three classifications of hearing loss level, hearing loss nature, and hearing pattern all used the following four features: the III-V wave interval, the V wave latency, the IV wave interval, and the subject's age.

[0050] In order to determine the importance of the four selected features, a feature ablation experiment was conducted. That is, on the basis of selecting all four features, a feature was removed each time for training. The drop in the accuracy of training after removing a feature represents the importance of this feature. The more the accuracy drops after removing this feature, the more important it is. The results of the feature ablation experiment after feature screening are shown as follows: Figure 3 shown.

[0051] Before proceeding to step S300, the waveform feature vectors and demographic feature vectors converted in step S200 are cleaned and standardized. Specifically, missing values and outliers are removed from the waveform feature vectors and demographic feature vectors, and the waveform feature vectors and demographic feature vectors are scaled to the interval [0, 1] to keep their eigenvalues within the same numerical range, facilitating learning for the neural network model. Finally, a set of cleaned and standardized waveform feature vectors and demographic feature vectors are output as input data for the auditory assessment model. The standardized waveform feature vectors and demographic feature vectors are denoted as x.

[0052] like Figure 1 As shown, the auditory assessment method based on fixed stimulation intensity further includes the following steps:

[0053] Step S300: Input the waveform feature vector and the demographic feature vector into a trained hearing assessment model, and output a hearing assessment result, wherein the hearing assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result.

[0054] Specifically, the standardized waveform feature vector and demographic feature vector are input into the trained auditory assessment model, and any one or more of the hearing loss level classification results, hearing loss nature classification results, and hearing graph classification results are output, thereby performing a more detailed mapping of the pure tone hearing threshold results from one or more dimensions of hearing loss level, hearing loss nature, and hearing graph.

[0055] Hearing loss level classification results include: normal hearing, mild hearing loss, moderate hearing loss, moderately severe hearing loss, and severe hearing loss. Hearing loss type classification results include: sensorineural, conductive, and mixed. Hearing pattern classification results include: flat, falling, rising, peak, and valley.

[0056] The embodiment of the present application performs auditory assessment through a neural network model, which effectively shortens the auditory assessment time, improves the assessment efficiency, and increases the feasibility of auditory assessment on large-scale populations. It can also obtain reliable and stable auditory assessment results based on fixed stimulation intensity.

[0057] In this embodiment of the present application, the training steps of the auditory assessment model include:

[0058] Obtaining a pre-constructed training set, the training set comprising a plurality of auditory brainstem response wave training data collected under a predetermined fixed stimulation intensity, and demographic characteristic training data, hearing loss level classification labels, hearing loss nature classification labels, and hearing pattern classification labels corresponding to each auditory brainstem response wave training data;

[0059] Extracting waveform feature training data corresponding to the auditory brainstem response wave training data, converting the waveform feature training data into waveform feature vector training data, and converting the demographic feature training data into demographic feature vector training data;

[0060] training a pre-built hearing loss degree classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss level classification labels; training a pre-built hearing loss property classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss property classification labels; and training a hearing pattern classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing pattern classification labels;

[0061] The trained hearing loss degree classification model, hearing loss nature classification model and hearing graph classification model are integrated into an auditory assessment model.

[0062] Specifically, the trained hearing loss degree classification model, hearing loss nature classification model and hearing graph classification model are integrated into the hearing assessment model, which means that the three models are connected in series to form a multi-dimensional judgment system at the application level, and each classification result is used to support decision-making in different dimensions. Figure 4 and Figure 5 As shown, Figure 4 and Figure 5 is a number of auditory brainstem response wave training data collected under a predetermined fixed stimulation intensity, that is, samples. Waveform features are extracted from the samples and converted into waveform feature vector training data, such as Figure 6 As shown; the waveform feature vector training data is filtered to obtain the standardized feature vector, as shown Figure 7 The neural network architectures of the hearing loss degree classification model, hearing loss nature classification model, and hearing pattern classification model are shown in Figure 1. Figure 8 shown.

[0063] The auditory assessment model in this embodiment is constructed based on the waveform characteristics of auditory brainstem response waves collected at a predetermined fixed stimulus intensity, thereby shortening the auditory assessment time and improving its efficiency. Furthermore, learning supervision is performed based on hearing loss level classification labels, hearing loss nature classification labels, and hearing pattern classification labels, achieving a more detailed mapping of pure tone hearing threshold results across the three dimensions of hearing loss level, hearing loss nature, and hearing pattern, thereby improving the comprehensiveness of the auditory assessment model.

[0064] In an embodiment of the present application, the hearing loss degree classification model, the hearing loss nature classification model, and the hearing graph classification model are all multi-layer feedforward neural networks. Specifically, the multi-layer feedforward neural network (Multi-Layer Perceptron, MLP) includes: an input layer, one or more hidden layers, and an output layer. The multi-layer feedforward neural network can learn complex nonlinear mapping relationships through nonlinear activation functions and multi-layer structures, and is widely used in tasks such as classification and regression. In addition to constructing an auditory assessment model based on a single model, the present application can also use a fusion of multiple model technologies to further improve the model performance.

[0065] The embodiment of the present application realizes nonlinear modeling and end-to-end learning by using a multi-layer feedforward neural network as an auditory assessment model, thereby improving the accuracy of auditory assessment.

[0066] In a specific embodiment, training a pre-built hearing loss classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss level classification label includes:

[0067] Inputting the waveform feature vector training data, the demographic feature vector training data, and the hearing loss level classification label into a pre-built hearing loss level classification model;

[0068] In the hearing loss degree classification model, the waveform feature vector training data and the demographic feature vector training data are processed by a linear transformation layer in the hearing loss degree classification model to obtain a linear transformation result;

[0069] Use nonlinear activation function to perform nonlinear mapping on the linear transformation result to obtain activation value;

[0070] The activation values are normalized and output using a random inactivation layer inserted between several hidden layers in the hearing loss degree classification model. The training is completed using the hearing loss level classification label as supervision to obtain a trained hearing loss degree classification model.

[0071] Training a pre-built hearing loss classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss classification label includes:

[0072] Inputting the waveform feature vector training data, the demographic feature vector training data, and the hearing loss nature classification label into a pre-built hearing loss nature classification model;

[0073] In the hearing loss classification model, the waveform feature vector training data and the demographic feature vector training data are processed by a linear transformation layer in the hearing loss classification model to obtain a linear transformation result;

[0074] Use nonlinear activation function to perform nonlinear mapping on the linear transformation result to obtain activation value;

[0075] The activation values are normalized and output using a random dropout layer inserted between several hidden layers in the hearing loss classification model. The training is completed using the hearing loss classification label as supervision to obtain a trained hearing loss classification model.

[0076] Training an auditory graph classification model based on the waveform feature vector training data, the demographic feature vector training data, and the auditory graph classification labels includes:

[0077] Inputting the waveform feature vector training data, the demographic feature vector training data, and the hearing graph classification labels into a pre-built hearing graph classification model;

[0078] In the hearing graph classification model, the waveform feature vector training data and the demographic feature vector training data are processed by a linear transformation layer in the hearing graph classification model to obtain a linear transformation result;

[0079] Use nonlinear activation function to perform nonlinear mapping on the linear transformation result to obtain activation value;

[0080] The activation values are normalized and output using a random dropout layer inserted between several hidden layers in the auditory graph classification model. The training is completed using the auditory graph classification label as supervision to obtain a trained auditory graph classification model.

[0081] Specifically, the waveform and demographic feature vector training data first pass through a linear transformation layer, performing multiplication and addition operations on the weight matrix and the input vector, achieving transformation and weighted combination of feature dimensions. The specific calculation formula is z = Wx + b; where W is the weight matrix, x is the standardized waveform and demographic feature vectors, b is the bias vector, and the output Z dimension is the dimension of the hidden layer.

[0082] After each linear transformation layer, a nonlinear activation function (ReLU) is applied to the linear output, transforming it nonlinearly and introducing nonlinear modeling capabilities. The specific formula is a = ReLU(z), where a is the activation value. The ReLU activation function maintains the dimension and only introduces nonlinear factors.

[0083] After the nonlinear activation function, the batch normalization operation is introduced to perform standardization processing on each batch of input data by returning the mean to zero and the variance to one, which helps to accelerate the convergence of the model and stabilize the training process.

[0084] In this embodiment, a random dropout layer is inserted between hidden layers to randomly block the output of some neurons according to a certain ratio to prevent overfitting of the model and enhance generalization ability. This process randomly sets the output of some neurons to 0 with a set probability p (such as 0.5), and the resulting dimension remains unchanged.

[0085] Specifically, the input data is processed through a linear transformation layer, a nonlinear activation function, batch normalization, and a random dropout layer. These steps are repeated several times before finally being connected to a fully connected output layer, which outputs the hearing status classification results based on the number of categories in the training task. The fully connected layer reduces the dimensionality of the data from the hidden layer to the required classification dimension (for example, for a five-category classification task, it is reduced to 5 dimensions).

[0086] In one embodiment of the present application, during the training process of the initial auditory assessment model, the training is performed with a preset cross entropy loss function as a target, and the model parameters are updated using an adaptive moment estimation optimizer.

[0087] Specifically, during the training process, the cross entropy loss function (for classification tasks) is used, and the adaptive moment estimation (Adam) optimizer is used to update the model parameters to improve the model performance. The model output is recorded as is the set of predicted probabilities for each category; for example, the categories of hearing loss levels include: normal hearing, mild hearing loss, moderate hearing loss, moderate to severe hearing loss, and severe hearing loss, then is the set of predicted probabilities for these five categories. The true label is y = [y1, y2, ..., y5]; the specific formula of the loss function is Where i is the category.

[0088] Furthermore, the training set is divided from the dataset. First, a dataset is obtained, which includes: a number of auditory brainstem response (ABSR) wave training data collected at a predetermined fixed stimulation intensity, along with demographic training data, hearing loss level classification labels, hearing loss nature classification labels, and hearing pattern classification labels corresponding to each ABSR wave training data. The dataset is divided into a training set and a validation set, and a 5-fold cross-validation method is used for model evaluation and parameter tuning to improve the model's generalization ability.

[0089] According to experimental data, the model constructed based on neural networks in this application has better classification efficiency in the three dimensions of hearing loss level, hearing loss nature and hearing graph than decision trees, random forests, Bayesian, logistic regression, support vector machines (SVM) and K-nearest neighbor algorithms (KNN), as shown in Table 1.

[0090] Table 1

[0091]

[0092] In one embodiment, if Figure 9 As shown, based on the above-mentioned auditory assessment method based on fixed stimulation intensity, the present invention also provides a corresponding auditory assessment device based on fixed stimulation intensity, comprising:

[0093] An acquisition module 100 is configured to acquire auditory brainstem response waves collected at a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves;

[0094] a conversion module 200, configured to extract waveform features of the auditory brainstem response wave, convert the waveform features into waveform feature vectors, and convert the demographic features into demographic feature vectors;

[0095] The output module 300 is used to input the waveform feature vector and the demographic feature vector into the trained hearing assessment model and output the hearing assessment results, which include: any one or more of the hearing loss level classification results, the hearing loss nature classification results and the hearing pattern classification results.

[0096] It should be noted that the above explanations of the embodiment of the hearing assessment method based on fixed stimulation intensity are also applicable to the hearing assessment device based on fixed stimulation intensity in this embodiment, and will not be repeated here.

[0097] The present invention discloses a hearing assessment device based on a fixed stimulation intensity. The device obtains auditory brainstem response waves collected at a predetermined fixed stimulation intensity and the demographic characteristics corresponding to the auditory brainstem response waves; extracts the waveform characteristics of the auditory brainstem response waves, converts the waveform characteristics into waveform feature vectors, and converts the demographic characteristics into demographic feature vectors; inputs the waveform feature vectors and demographic feature vectors into a trained hearing assessment model, and outputs hearing assessment results. The hearing assessment results include any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result. By processing auditory brainstem response data collected at a predetermined fixed stimulation intensity, the present application reduces the number of tests required to assess hearing level, thereby shortening the hearing assessment time. Furthermore, by outputting recognition results of one or more dimensions of the hearing loss level classification result, the hearing loss nature classification result, and the hearing pattern classification result based on the hearing assessment model, the accuracy of the hearing assessment is improved.

[0098] Figure 10 This is a schematic diagram of the structure of a terminal provided in an embodiment of the present application. The terminal may include:

[0099] Memory 501 , processor 502 , and computer programs stored in the memory 501 and executable on the processor 502 .

[0100] When the processor 502 executes the program, the auditory assessment method based on fixed stimulation intensity provided in the above embodiment is implemented.

[0101] Furthermore, the terminal further includes:

[0102] The communication interface 503 is used for communication between the memory 501 and the processor 502 .

[0103] The memory 501 is used to store computer programs that can be run on the processor 502 .

[0104] The memory 501 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0105] If the memory 501, processor 502, and communication interface 503 are implemented independently, the communication interface 503, memory 501, and processor 502 can be interconnected via a bus to enable communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, control buses, and the like. For ease of illustration, the figure shows only one line, but this does not imply that there is only one bus or only one type of bus.

[0106] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0107] The processor 502 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0108] This embodiment further provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the above-mentioned auditory assessment method based on fixed stimulation intensity is implemented.

[0109] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0110] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0111] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing a custom logical function or process step, and the scope of the preferred embodiments of the present application includes additional implementations in which the order shown or discussed may not be followed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by a person skilled in the art to which the embodiments of the present application belong.

[0112] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can read and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or N wires (electronic devices), a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting, or otherwise processing in a suitable manner as necessary, and then storing it in a computer memory.

[0113] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0114] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above-mentioned embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0115] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0116] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A method for auditory assessment based on fixed stimulus intensity, characterized in that: The method comprises: obtaining auditory brainstem response waves collected at a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves; extracting waveform features of the auditory brainstem response wave, converting the waveform features into waveform feature vectors, and converting the demographic features into demographic feature vectors; The waveform feature vector and the demographic feature vector are input into a trained hearing assessment model to output a hearing assessment result, wherein the hearing assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result.

2. The auditory assessment method based on fixed stimulation intensity according to claim 1, characterized in that: The predetermined fixed stimulation intensity is 80 dB nHL.

3. The auditory assessment method based on fixed stimulation intensity according to claim 1, characterized in that: The waveform characteristics include: one or more of: the latency of the predetermined reaction wave, the amplitude corresponding to the predetermined reaction wave, the interwave interval between predetermined reaction waves, and the amplitude ratio between predetermined reaction waves; the demographic characteristics include: the gender and / or age of the subject.

4. The auditory assessment method based on fixed stimulation intensity according to claim 1, characterized in that: The training steps of the auditory evaluation model include: Obtaining a pre-constructed training set, the training set comprising a plurality of auditory brainstem response wave training data collected under a predetermined fixed stimulation intensity, and demographic characteristic training data, hearing loss level classification labels, hearing loss nature classification labels, and hearing pattern classification labels corresponding to each auditory brainstem response wave training data; Extracting waveform feature training data corresponding to the auditory brainstem response wave training data, converting the waveform feature training data into waveform feature vector training data, and converting the demographic feature training data into demographic feature vector training data; training a pre-built hearing loss degree classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss level classification labels; training a pre-built hearing loss nature classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing loss nature classification labels; and training a hearing pattern classification model based on the waveform feature vector training data, the demographic feature vector training data, and the hearing pattern classification labels; The trained hearing loss degree classification model, hearing loss nature classification model and hearing graph classification model are integrated into an auditory assessment model.

5. The auditory assessment method based on fixed stimulation intensity according to claim 4, characterized in that: The hearing loss degree classification model, the hearing loss nature classification model and the hearing pattern classification model are all multi-layer feedforward neural networks.

6. The auditory assessment method based on fixed stimulation intensity according to claim 5, characterized in that: The hearing loss degree classification model and the hearing loss nature classification model both adopt a four-layer neural network architecture, and the dimension of the hidden layer of the hearing loss degree classification model and the hearing loss nature classification model is 128 dimensions; the hearing graph classification model adopts a five-layer neural network architecture, and the dimension of the hidden layer of the hearing graph classification model is 256 dimensions.

7. The auditory assessment method based on fixed stimulation intensity according to claim 5, characterized in that: During the training process of the hearing loss degree classification model, the hearing loss nature classification model, and the hearing pattern classification model, a cross entropy loss function is used for classification tasks, and an adaptive moment estimation optimizer is used to update the model parameters.

8. An auditory assessment device based on fixed stimulation intensity, characterized in that: The device comprises: an acquisition module, configured to acquire auditory brainstem response waves collected under a predetermined fixed stimulation intensity and demographic characteristics corresponding to the auditory brainstem response waves; a conversion module, configured to extract waveform features of the auditory brainstem response wave, convert the waveform features into waveform feature vectors, and convert the demographic features into demographic feature vectors; An output module is used to input the waveform feature vector and the demographic feature vector into a trained hearing assessment model and output a hearing assessment result, wherein the hearing assessment result includes any one or more of a hearing loss level classification result, a hearing loss nature classification result, and a hearing pattern classification result.

9. A terminal, characterized in that: include: A memory, a processor, and a hearing assessment program based on fixed stimulation intensity stored in the memory and executable on the processor, wherein the hearing assessment program based on fixed stimulation intensity, when executed by the processor, implements the steps of the hearing assessment method based on fixed stimulation intensity according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which can be executed to implement the steps of the auditory assessment method based on fixed stimulation intensity according to any one of claims 1 to 7.

Images