Hearing aid intelligent protection device integrated on headgear and method

Abstract

The invention provides a hearing aid intelligent protection device and method integrated on a headgear, and the method comprises the steps: converting an environment sound wave signal into a frequency domain signal through a signal processor in an interlayer of the headgear, and separating a noise dominant frame and a voice dominant frame in a hearing aid; determining a voice masking index between the environmental noise and the voice in the environmental sound wave signal according to a distribution association relationship between the noise dominant frame and the voice dominant frame; according to the energy distribution characteristics of the noise spectrum in the frequency domain signal, determining the noise interference suffered by the hearing aid in the current wearing scene, and determining the transient noise dynamic characteristics of the environment sound wave signal under the current noise interference; and performing disturbance trend prediction on voice signal noise in the hearing aid according to the transient noise dynamic characteristics and the voice masking index, and further updating an adaptive filtering coefficient of the hearing aid according to a prediction result. By adopting the scheme of the invention, the prediction of the noise disturbance trend in a complex and changeable environment can be realized so as to adaptively adjust the noise reduction filter coefficient of the hearing aid.

Description

Technical Field

[0001] The present application relates to the technical field of hearing aids, and more specifically, to an intelligent hearing aid protection device and method integrated on a headgear. Background Art

[0002] A hearing aid is an electronic device used to improve the hearing ability of hearing-impaired patients. Its core function is to amplify voice signals and suppress environmental noise to enhance the user's voice clarity and hearing experience in complex acoustic environments. With the development of technology, hearing aids are gradually integrated with wearable devices (such as headphones, smart hats, etc.) to meet users' needs for concealment, comfort and scene adaptability. Hearing aids, as hearing-assistive devices for people with hearing loss, are widely used in daily communication, work and life assistance for the elderly, hearing-impaired patients and hearing-sensitive people to enhance users' perception of voice signals.

[0003] Intelligent hearing aid protection integrated into a headband can reduce the noise of the hearing aid's voice signal, meeting the hearing protection and enhancement needs of the elderly and hearing-impaired people in complex scenarios such as daily travel and work communication. However, existing hearing aid noise reduction technologies mostly rely on fixed noise reduction filter coefficient settings or simple environmental sound detection methods. On the one hand, traditional hearing aids find it difficult to accurately identify the coupling state of environmental noise and voice signals in real time based on complex and changing wearing scenarios, such as noisy streets and raucous indoor gatherings. On the other hand, when the environmental noise suddenly changes, such as a sudden car horn or machine roar, existing noise reduction methods cannot quickly and accurately capture the signal mutation under noise interference, which leads to inaccurate noise processing in the voice signal, ultimately resulting in poor noise reduction effect and unable to meet users' needs for clear hearing in diverse scenarios. Therefore, how to predict the noise disturbance trend in complex and changing environments to adaptively adjust the noise reduction filter coefficient of the hearing aid has become a difficult problem facing the industry. Summary of the Invention

[0004] The present application provides an intelligent hearing aid protection device and method integrated on a head cap, which can predict the noise disturbance trend in a complex and changeable environment to adaptively adjust the noise reduction filter coefficient of the hearing aid.

[0005] In a first aspect, the present application provides a hearing aid adaptive noise reduction method for a hearing aid, wherein an intelligent hearing aid protective device integrated on a headgear performs adaptive noise reduction on the hearing aid, wherein the intelligent hearing aid protective device integrated on the headgear includes an intelligent sensor and a signal processor, and the method comprises the following steps: The intelligent sensors integrated in the corresponding positions of the ears on both sides of the head cap collect the environmental sound wave signals, and transmit the collected environmental sound wave signals to the signal processor in the interlayer of the head cap; The signal processor converts the ambient sound wave signal into a frequency domain signal, and then separates the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal; Determining a speech masking index between ambient noise and speech in the ambient sound wave signal based on a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; Determining the noise interference experienced by the hearing aid in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and determining the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the mutation amplitude of the noise in the noise interference; The disturbance trend of the speech signal noise in the hearing aid is predicted based on the transient noise dynamic characteristics and the speech masking index, and the adaptive filter coefficient of the hearing aid is updated according to the prediction result.

[0006] In some embodiments, converting the ambient sound wave signal into a frequency domain signal by the signal processor specifically includes: The signal processor divides the ambient sound wave signal into frames to obtain a plurality of audio frames; Perform frequency domain conversion on each audio frame to obtain a frequency domain signal of the ambient sound wave signal.

[0007] In some embodiments, separating the noise-dominated frames and the speech-dominated frames in the hearing aid based on the noise characteristics and the speech characteristics in the frequency domain signal specifically includes: extracting noise features and speech features from the frequency domain signal; matching a noise-dominant frame in the hearing aid from the frequency domain signal according to the noise characteristics; A speech dominant frame in the hearing aid is matched from the frequency domain signal based on the speech feature.

[0008] In some embodiments, extracting noise features and speech features from the frequency domain signal specifically includes: Determining a spectral entropy value of each audio frame in the frequency domain signal; determining a speech band distribution of the frequency domain signal based on all spectral entropy values; Noise features and speech features are extracted from the frequency domain signal according to the speech band distribution.

[0009] In some embodiments, determining the speech masking index between the ambient noise and the speech in the ambient sound wave signal based on the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames specifically includes: Determining a ratio distribution diagram of the noise-dominated frames and the speech-dominated frames; extracting a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames from the proportion distribution graph; Determining an inter-frame overlap between the noise-dominated frame and the speech-dominated frame based on the distribution association relationship; The speech masking index between the ambient noise and the speech in the ambient sound wave signal is determined according to the inter-frame overlap.

[0010] In some embodiments, determining the noise interference to which the hearing aid is subjected in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal specifically includes: determining energy distribution characteristics of a noise spectrum in the frequency domain signal; determining a noise classification index of the frequency domain signal based on the energy distribution characteristics; The noise interference to which the hearing aid is subjected in the current wearing scenario is determined by the noise classification index.

[0011] In some embodiments, predicting the disturbance trend of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index specifically includes: training a trend prediction model of speech signal noise disturbance in a hearing aid based on the transient noise dynamic characteristics and the speech masking index; The disturbance trend of the speech signal noise in the hearing aid is predicted using the trained trend prediction model.

[0012] In a second aspect, the present application provides an intelligent hearing aid protective device integrated into a headgear, including an adaptive noise reduction unit, wherein the adaptive noise reduction unit includes: The acquisition module is used to collect ambient sound wave signals through smart sensors integrated in the corresponding positions of the ears on both sides of the headwear, and transmit the collected ambient sound wave signals to the signal processor in the interlayer of the headwear; a processing module, configured to convert the ambient sound wave signal into a frequency domain signal by the signal processor, and then separate the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal; The processing module is further configured to determine a speech masking index between the ambient noise and the speech in the ambient sound wave signal based on a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; The processing module is further configured to determine the noise interference experienced by the hearing aid in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and determine the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the mutation amplitude of the noise in the noise interference; An execution module is used to predict the disturbance trend of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index, and then update the adaptive filter coefficient of the hearing aid according to the prediction result.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device performs the above-mentioned hearing aid adaptive noise reduction method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions or codes. When the instructions or codes are executed on a computer, the computer implements the above-mentioned hearing aid adaptive noise reduction method.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In the present application, ambient sound wave signals are collected by intelligent sensors integrated in the corresponding positions of the ears on both sides of the headgear, and the collected ambient sound wave signals are transmitted to a signal processor in the interlayer of the headgear; the signal processor converts the ambient sound wave signals into frequency domain signals, and then separates the noise-dominated frames and speech-dominated frames in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signals; the speech masking index between the ambient noise and speech in the ambient sound wave signal is determined based on the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; the noise interference suffered by the hearing aid in the current wearing scenario is determined based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference are determined based on the mutation amplitude of the noise in the noise interference; the disturbance trend of the speech signal noise in the hearing aid is predicted based on the transient noise dynamic characteristics and the speech masking index, and the adaptive filter coefficient of the hearing aid is updated based on the prediction result.

[0016] It can be seen that in the present application, firstly, the speech masking index between the ambient noise and the speech in the ambient sound wave signal is determined based on the distribution correlation relationship between the noise-dominated frame and the speech-dominated frame, which can accurately characterize the masking degree and interference sensitivity of the speech in the noise interference, and lay a context-dependent basis for the subsequent determination of whether the noise causes continuous or sudden disturbance to the speech, and can improve the hearing aid's perception of complex speech-noise relationships in real scenes; secondly, the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference are determined based on the mutation amplitude of the noise in the noise interference, which can improve the quantification ability of the noise interference intensity and change rate, and can avoid speech masking in the hearing aid or user hearing burden caused by sudden high-intensity interference, thereby preparing noise reduction adjustment in advance and providing necessary data support for subsequent noise disturbance trend prediction; then, the transient noise dynamic characteristics are determined based on the mutation amplitude of the noise in the noise interference, which can improve the quantification ability of the noise interference intensity and change rate, and can avoid speech masking in the hearing aid or user hearing burden caused by sudden high-intensity interference, thereby preparing noise reduction adjustment in advance and providing necessary data support for subsequent noise disturbance trend prediction. The dynamic characteristics of the state noise and the speech masking index are used to predict the disturbance trend of the speech signal noise in the hearing aid, so that the hearing aid can have the ability to perceive the future noise interference dynamics in advance, and no longer rely solely on the current state for passive response, which significantly improves the forward-looking adaptability of the hearing aid noise reduction process to complex environments; finally, the adaptive filter coefficient of the hearing aid is updated according to the prediction results, so that the hearing aid can adapt quickly when facing sudden or changeable noise, effectively reducing the noise interference of the speech signal in the hearing aid, while avoiding the weakening of speech quality due to excessive noise reduction due to sudden intervention, improving the noise reduction efficiency and speech clarity of the hearing aid, so that the hearing aid can achieve dynamic optimal hearing adjustment in different usage scenarios, so as to enhance the auditory comfort and communication efficiency of the hearing aid user; in summary, this scheme can realize the prediction of noise disturbance trends in complex and changeable environments to adaptively adjust the noise reduction filter coefficient of the hearing aid. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0018] Figure 1 is an exemplary flow chart of a hearing aid adaptive noise reduction method according to some embodiments of the present application; Figure 2 is an exemplary flow chart for implementing frame division according to some embodiments of the present application; Figure 3 is an exemplary flow chart of determining noise interference according to some embodiments of the present application; Figure 4 is a schematic structural diagram of an adaptive noise reduction unit according to some embodiments of the present application; Figure 5 2 is a schematic diagram of the structure of a computer device for implementing a hearing aid adaptive noise reduction method according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0020] refer to Figure 1 , which is an exemplary flow chart of a hearing aid adaptive noise reduction method according to some embodiments of the present application. The hearing aid adaptive noise reduction method 100 mainly includes the following steps: In step 101, ambient sound wave signals are collected by intelligent sensors integrated in the corresponding positions of the ears on both sides of the headgear, and the collected ambient sound wave signals are transmitted to a signal processor in the interlayer of the headgear.

[0021] In a specific implementation, the collection of ambient sound wave signals by smart sensors integrated in the corresponding positions of the ears on both sides of the headgear can be achieved in the following manner, namely, a miniature smart sensor (i.e., a microphone module with signal sensing and preliminary processing capabilities) can be embedded in the corresponding positions of the left and right ears of the headgear. The smart sensor samples the sound wave changes in the surrounding air in real time through a sound pressure sensing method, converts the sound wave signal into a voltage signal after sampling, and processes the analog sound wave signal into stable and interference-resistant digital audio data through a built-in analog preamplifier, bandpass filtering, and analog-to-digital conversion module, thereby obtaining the ambient sound wave signal; the collected ambient sound wave signal can be transmitted to the signal processor in the interlayer of the headgear in the following manner, namely, the collected ambient sound wave signal can be synchronously transmitted to the signal processor in the interlayer of the headgear through a serial audio communication (Inter-IC Sound, I²S) or a serial peripheral interface (Serial Peripheral Interface, SPI) communication protocol for subsequent frequency domain conversion and noise reduction analysis; other methods can also be used in other embodiments, which are not specifically limited here.

[0022] It should be noted that the smart sensor in this application is a microphone module for capturing ambient sounds, which is used to capture sound waves in the left and right ears of the hearing aid user; the signal processor in this application is the core hardware unit for performing subsequent signal analysis and feature extraction, which is integrated in a low-power digital signal processing chip (DSP) or system-on-chip (SoC) in the interlayer of the headband; the ambient sound wave signal in this application represents the ambient sound signal initially captured by the hearing aid, which serves as the original data source for noise reduction processing by the hearing aid.

[0023] In step 102, the signal processor converts the ambient sound wave signal into a frequency domain signal, and then separates the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal.

[0024] In some embodiments, the signal processor converts the ambient sound wave signal into a frequency domain signal by using the following steps: The signal processor divides the ambient sound wave signal into frames to obtain a plurality of audio frames; Perform frequency domain conversion on each audio frame to obtain a frequency domain signal of the ambient sound wave signal.

[0025] For specific implementation, refer to Figure 2 As shown, this figure is an exemplary flow chart for implementing frame division in some embodiments of the present application. The signal processor performs frame division on the ambient sound wave signal to obtain multiple audio frames, which can be implemented in the following manner: the signal processor performs frame division on the collected ambient sound wave signal, and the continuous time domain signal of the ambient sound wave signal can be divided into a series of short-time stable audio frames by setting a sliding window of fixed length on the time axis (such as 20ms window length, 10ms overlap length). This process is often weighted based on a Hamming window, a rectangular window or a Hanning window function to avoid spectrum leakage at the frame boundary and maintain the stability of the short-time signal, thereby providing a basis for subsequent frequency domain analysis; frequency domain conversion is performed on each audio frame to obtain the frequency domain signal of the ambient sound wave signal, which can be implemented in the following manner, namely: Fast Fourier transform (Fast Fourier transform) is applied to each audio frame. Transform (FFT) converts the time domain samples in the audio frame into their frequency components to obtain a frequency domain signal containing the amplitude and phase information of each frequency component, thereby obtaining the frequency domain signal of the ambient sound wave signal to provide information about the energy distribution of the ambient sound at different frequencies, thereby providing a physical basis for the subsequent extraction of noise features and speech features; other methods may also be used for determination in other embodiments, which are not limited here.

[0026] It should be noted that the audio frame in this application represents a short-time continuous signal divided in time from the ambient sound wave signal, which has local stationarity and is conducive to performing short-time frequency domain analysis; the frequency domain signal in this application represents the energy distribution signal of the ambient sound received by the hearing aid at different frequencies, through which the distribution pattern of noise and speech in the hearing aid in the spectrum can be identified.

[0027] In some embodiments, separating the noise-dominated frames and the speech-dominated frames in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal can be achieved by using the following steps: extracting noise features and speech features from the frequency domain signal; matching a noise-dominant frame in the hearing aid from the frequency domain signal according to the noise characteristics; A speech dominant frame in the hearing aid is matched from the frequency domain signal based on the speech feature.

[0028] In some embodiments, extracting noise features and speech features from the frequency domain signal may be achieved by using the following steps: Determining a spectral entropy value of each audio frame in the frequency domain signal; determining a speech band distribution of the frequency domain signal based on all spectral entropy values; Noise features and speech features are extracted from the frequency domain signal according to the speech band distribution.

[0029] In specific implementation, determining the spectral entropy value of each audio frame in the frequency domain signal can be achieved in the following manner, namely: for the frequency domain representation of each audio frame in the frequency domain signal, first calculate its normalized spectral energy distribution, that is, normalize the power of each frequency component to a percentage of the total energy; then use the Shannon entropy formula to calculate the information entropy of the normalized spectral energy distribution, and use the obtained calculation result as the spectral entropy value of each audio frame, which reflects the degree of dispersion of the spectrum energy of each audio frame in the frequency domain signal. A high spectral entropy value indicates uniform energy distribution (typically a noise frame), and a low spectral entropy value indicates concentrated energy (typically a speech frame); determining the speech band distribution of the frequency domain signal based on all spectral entropy values ​​can be achieved in the following manner, namely: performing statistical analysis on all spectral entropy values, and preliminarily classifying the audio frames corresponding to the spectral entropy values ​​by setting an entropy value threshold (for example, setting by empirical method or grouping by K-means clustering method), that is, classifying the audio frames with low spectral entropy values. By summarizing and analyzing the frequency components corresponding to the audio frames, the frequency interval where the energy of the frequency domain signal is primarily concentrated, i.e., the speech band distribution, can be obtained. This represents the frequency range in which the speech signal is typically focused in the frequency domain, typically 300 Hz to 3400 Hz or a dynamically varying range thereof. Extracting noise features and speech features from the frequency domain signal based on the speech band distribution can be achieved by performing power spectrum estimation on the frequency components of the frequency domain signal that fall within and outside the speech band distribution, respectively, to extract their statistical parameters (e.g., energy distribution, volatility, peak amplitude, etc.). Features that are periodically concentrated within the speech band distribution are considered speech features, while features that are diffuse and lack a periodic structure in the non-speech band distribution are classified as noise features for subsequent frame classification and adaptive filter coefficient adjustment. Other methods may also be used for determination in other embodiments, which are not limited here.

[0030] It should be noted that the spectral entropy value in this application represents the degree of dispersion of the spectral energy of the audio frame in the frequency domain signal, which is a structural characteristic used to distinguish between speech frames and noise frames in the frequency domain signal; the voice band distribution in this application represents the main frequency band in the frequency domain signal where energy is concentrated to reflect speech activity, which is a key reference for identifying speech components; the noise feature in this application represents the feature of the non-structural energy diffusion component in the frequency domain signal, through which noise interference signals can be identified; the speech feature in this application represents the feature of the periodic and concentrated energy distribution in the frequency domain signal, which will not be repeated here.

[0031] In a specific implementation, matching the noise-dominant frames in the hearing aid from the frequency domain signal based on the noise characteristics can be achieved in the following manner, namely, the spectrum of each audio frame in the frequency domain signal can be matched with the noise characteristics (such as high spectral diffusion, large spectral entropy value, sudden increase in low-frequency intensity, etc.) for similarity, and then a classification algorithm (such as a threshold method, a support vector machine, or a K-Nearest Neighbors (KNN) algorithm) can be used to select audio frames whose matching results exceed the threshold as noise-dominant frames in the hearing aid. The noise-dominant frames generally refer to frames whose spectral characteristics are most similar to known noise forms (such as wind noise and background noise) and whose speech feature distribution is not obvious, which will not be described in detail here. Matching the speech-dominant frames in the hearing aid from the frequency domain signal based on the speech features can be achieved in the following manner, namely, comparing the spectrum of each audio frame in the frequency domain signal with the speech characteristics, such as whether there is a clear fundamental frequency peak, the spectrum is concentrated in the voice band range, and the spectral entropy value is low, and then using voice activity detection (Voice Activity Detection) to determine whether the frame has a clear fundamental frequency peak, the spectrum is concentrated in the voice band range, and the spectral entropy value is low. A VAD) model or a threshold-based rule classification algorithm is used to identify frames that dominate speech features from the compared audio frames, and use them as the speech-dominant frames in the hearing aid to capture the portion of the frequency domain signal that reflects a higher proportion of real speech components and is more meaningful for hearing aid gain adjustment. Other methods may also be used in other embodiments, which are not limited here.

[0032] It should be noted that the noise-dominated frame in this application refers to a signal frame whose main energy source is the environmental noise component; the speech-dominated frame in this application refers to a signal frame whose main energy source is the speech component, so effective language information can be identified through it.

[0033] In step 103, a speech masking index between the ambient noise and the speech in the ambient sound wave signal is determined according to the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames.

[0034] In some embodiments, determining the speech masking index between the ambient noise and the speech in the ambient sound wave signal based on the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames can be achieved by using the following steps: Determining a ratio distribution diagram of the noise-dominated frames and the speech-dominated frames; extracting a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames from the proportion distribution graph; Determining an inter-frame overlap between the noise-dominated frame and the speech-dominated frame based on the distribution association relationship; The speech masking index between the ambient noise and the speech in the ambient sound wave signal is determined according to the inter-frame overlap.

[0035] In a specific implementation, determining the ratio distribution map of the noise-dominated frames and the speech-dominated frames can be achieved in the following manner, namely, dividing the audio frames of the ambient sound wave signal in the time series into time windows of equal length (e.g., 100 ms per group), counting the number of noise-dominated frames and speech-dominated frames identified in each window, calculating the respective proportions of noise-dominated frames and speech-dominated frames, and plotting them to form a ratio distribution map to visualize the relative density of the two types of frames on the time axis, which serves as a basis for analyzing the temporal relationship of the frames; extracting the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames from the ratio distribution map can be achieved in the following manner, namely, using a sliding window analysis method to extract the temporal correlation between the speech-dominated frames and the noise-dominated frames from the ratio distribution map, for example, counting whether the two types of frames frequently co-occur in the same time window, or whether they show a sequence pattern (e.g., noise precedes speech bursts), and using the statistical structure obtained by describing whether the two types of frames have a synchronous, alternating, or independent statistical structure on the time axis as the distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; in other embodiments, other methods can also be used for determination, which is not limited here.

[0036] In a specific implementation, the inter-frame overlap between the noise-dominated frame and the speech-dominated frame can be determined by the distribution association relationship in the following manner, namely, by sliding matching or dynamic time warping. A time warping (DTW) algorithm is used to calculate the intersection of the occurrence intervals of noise-dominated frames and speech-dominated frames in a time series, and the calculated result is used as the inter-frame overlap between the noise-dominated frames and the speech-dominated frames. The speech masking index between the ambient noise and speech in the ambient sound wave signal is determined by the inter-frame overlap, which can be implemented in the following manner: the coupling degree between the ambient noise and speech in the ambient sound wave signal can be evaluated based on an inter-frame overlap threshold (e.g., coupling is considered to exist when the overlap exceeds 30%), and the evaluation result is used as the speech masking index between the ambient noise and speech in the ambient sound wave signal, for example, weak coupling masking (speech and ambient noise are essentially independent in time), medium coupling masking (partial co-occurrence), and strong coupling masking (speech and noise frequently alternate or mix in the same frame), to determine whether there is obvious noise interference in the speech signal, and provide a basis for the subsequent adaptive selection of core parameters for the noise reduction mode. For example, in the strong coupling masking state, the speech retention ability needs to be improved, and in the weak coupling masking state, the background noise suppression strength can be enhanced. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0037] It should be noted that the proportional distribution diagram in this application is used to quantify the time distribution differences and overlaps between noise-dominated frames and speech-dominated frames; the distribution correlation relationship in this application represents the correlation law between the time distribution of noise-dominated frames and speech-dominated frames, through which it can be judged whether there may be coupling interference between the two; the inter-frame overlap in this application represents the proportion of overlapping occurrences of noise-dominated frames and speech-dominated frames in the same time segment, which is a quantitative basis for judging the degree of signal overlap; the speech masking index in this application represents the degree of overlapping mixing between ambient noise and speech signals in ambient sound wave signals, through which the subsequent adaptive noise reduction adjustment effect of hearing aids can be precisely controlled.

[0038] In step 104, the noise interference to the hearing aid in the current wearing scenario is determined based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference are determined based on the sudden change amplitude of the noise in the noise interference.

[0039] In some embodiments, reference Figure 3 As shown in FIG, this figure is an exemplary flow chart for determining noise interference in some embodiments of the present application. In this embodiment, the noise interference to the hearing aid in the current wearing scenario is determined based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, which can be achieved by the following steps: First, in step 1041, the energy distribution characteristics of the noise spectrum in the frequency domain signal are determined; Next, in step 1042, the noise level index of the frequency domain signal is determined based on the energy distribution characteristics; Finally, in step 1043, the noise interference to which the hearing aid is subjected in the current wearing scenario is determined using the noise classification index.

[0040] In a specific implementation, the energy distribution characteristics of the noise spectrum in the frequency domain signal can be determined by the following methods: using short-time Fourier transform to identify the non-speech frequency band in the frequency domain signal and extracting the energy distribution characteristics therein, such as spectral centroid (describing the frequency band where noise energy is concentrated), spectral variance (reflecting the degree of noise energy diffusion), spectral flatness (measuring the complexity of noise energy), etc.; determining the noise classification index of the frequency domain signal from the energy distribution characteristics can be achieved by the following methods: inputting the energy distribution characteristics into a trained noise classification model (such as a noise classification model based on a support vector machine (SVM)). A noise classification model is used to output a noise classification index (between 1 and 10) to reflect the intensity level of the ambient noise in the spectrum signal. Taking a support vector machine (SVM) as an example, the noise classification model is trained by first collecting a large number of frequency domain signal samples under different noise environments and extracting their energy distribution feature vectors as input features. Each sample is labeled to determine its corresponding noise level label (e.g., 1 to 10). The sample data is divided into a training set and a test set. The SVM model is trained using the training set. By selecting an appropriate kernel function (e.g., a radial basis function kernel) and adjusting the model's hyperparameters (e.g., a penalty parameter C), the model learns the mapping relationship between the input features and the noise level labels. The model's performance is then verified and optimized using the test set, ultimately resulting in a trained SVM noise classification model. In other embodiments, the noise classification model may also be trained using other methods, which are not limited here.

[0041] In specific implementation, the noise interference to which the hearing aid is subjected in the current wearing scenario can be determined by the noise grading index in the following manner, namely: first, a noise interference threshold can be set according to the noise grading index and a large amount of historical experimental data, wherein the noise interference threshold is used to extract the noise interference in the hearing aid. In addition, the noise interference threshold is set in a manner such as: during the configuration stage of the hearing aid, a large amount of historical experimental data is collected, including ambient sound wave signal samples in different noise environments, and the corresponding noise grading index, and each ambient sound wave signal sample must be converted into the frequency domain to obtain its spectral energy distribution (such as spectral centroid, variance, etc.), and then the spectral energy distribution of the sample corresponding to each noise grading index is statistically analyzed to determine the noise level in the noise grading index. Under the grading index, the energy range of the frequency signal of the noise interference is obtained. Then, for each noise grading index, a noise interference threshold is set. The noise interference threshold can effectively distinguish the frequency energy threshold of the noise and the target speech signal, and the noise interference threshold corresponding to the noise grading index can be found; then, the noise interference threshold can be used to fuse the spectral energy distribution of the ambient sound wave signal to construct a binary mask matrix, and then the ambient sound wave signal in the hearing aid is multiplied based on the binary mask matrix to extract a set of noise sub-segments with obvious noise energy characteristics in the signal received by the hearing aid as the noise interference received by the hearing aid in the current wearing scenario. Other methods can also be used to implement this in other embodiments, which is not limited here.

[0042] It should be noted that the energy distribution characteristics in this application represent the specific distribution of noise energy in the ambient sound wave signal; the noise grading index in this application represents the grading index of the strength of ambient noise in the spectrum signal; the noise interference in this application represents the non-speech sound signal that interferes with speech recognition and auditory comfort in the user wearing scenario, through which the acceptable background sound and the noise that needs to be processed in the hearing aid can be distinguished, providing a criterion for the noise reduction target.

[0043] In some embodiments, determining the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the sudden change amplitude of the noise in the noise interference can be achieved by using the following steps: Determining a sudden change amplitude of noise in the noise interference; Extracting a plurality of noise mutation frames in the ambient sound wave signal based on the mutation amplitude; The transient noise dynamic characteristics of the environmental sound wave signal under noise interference are determined through all noise mutation frames.

[0044] In a specific implementation, determining the mutation amplitude of the noise in the noise interference can be achieved in the following manner, namely, performing a time difference analysis on the noise interference, such as calculating the change value of the total energy (such as the root mean square energy (RMS)) and the spectral centroid position between each noise sub-segment in the noise interference and its adjacent previous noise sub-segment, obtaining the energy mutation value and the spectral centroid change between adjacent noise sub-segments in the noise interference, and all the energy mutation values ​​and spectral centroid changes constitute the mutation amplitude of the noise in the noise interference; extracting multiple noise mutation frames in the ambient sound wave signal from the mutation amplitude can be achieved in the following manner, namely, setting a threshold (such as energy surge > 6 dB, spectral centroid change > 300 Hz) in combination with an empirical threshold to filter out all noise sub-segments with mutation amplitudes exceeding the set threshold, and performing time clustering processing on the filtered noise sub-segments, that is, merging noise sub-segments that are continuous or have a time interval of less than 50 ms into a noise mutation frame, and excluding isolated single frames (less than two consecutive frames) or pseudo mutation frames with a duration of too short (less than 100 ms) to retain valid noise mutations. Frames can be used to obtain multiple noise mutation frames in the ambient sound wave signal, so as to avoid misjudging false mutations caused by occasional spikes and effectively extract real sudden noise events; determining the transient noise dynamic characteristics of the ambient sound wave signal under noise interference through all noise mutation frames can be achieved in the following way, namely: the distribution density (such as mutation frequency), average interval, duration and energy peak of all noise mutation frames on the time axis can be statistically analyzed, and then the transient noise dynamic characteristics of the ambient sound wave signal, such as "average mutation frequency", "mutation intensity distribution", "spectrum deviation trend" and other dynamic characteristic parameters, are constructed based on the obtained statistical results to describe the activity level and evolution trend of the mutation noise in the current wearing scenario, which can be used as the data basis for subsequent adaptive filter coefficient adjustment; in other embodiments, other methods can also be used for determination, which is not limited here.

[0045] It should be noted that the mutation amplitude in this application represents the degree of change of noise in noise interference, which can be used to identify noise mutation behavior in ambient sound wave signals; the noise mutation frame in this application represents a signal frame in which noise in ambient sound wave signals changes rapidly, which is usually highly correlated with external sudden noise (such as impact sound, horn sound); the transient noise dynamic characteristics in this application represent the changing characteristics of the mutation noise interference in the current sound environment of the hearing aid, which is the key basis for adaptive noise control of the hearing aid.

[0046] In step 105, the disturbance trend of the speech signal noise in the hearing aid is predicted based on the transient noise dynamic characteristics and the speech masking index, and the adaptive filter coefficient of the hearing aid is updated according to the prediction result.

[0047] In some embodiments, the disturbance trend prediction of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index can be achieved by using the following steps: training a trend prediction model of speech signal noise disturbance in a hearing aid based on the transient noise dynamic characteristics and the speech masking index; The disturbance trend of the speech signal noise in the hearing aid is predicted using the trained trend prediction model.

[0048] It should be noted that the trend prediction model in this application refers to a neural network structure used to predict the noise disturbance trend of the speech signal in the hearing aid, and its function is to provide a feedforward adjustment basis for the noise reduction of the hearing aid. The trend prediction model is constructed based on existing labeled speech-noise sample data through supervised learning. During the training process, the neural network model learns the nonlinear mapping relationship between the dynamic characteristics of transient noise in the environmental sound wave signal and the speech masking index and the subsequent speech noise interference degree. The input data is supervised and trained. The long short-term memory network (LSTM) or Transformer structure is used to model the time dependence of the input features, and the historical speech quality change trend marked in the training sample is used as a supervision signal. The network parameters are optimized through backpropagation and multiple rounds of gradient descent to minimize the prediction error. The trend prediction model can output a predicted value of the noise disturbance degree of the speech signal in the hearing aid, thereby realizing dynamic feedforward control of the hearing aid noise reduction strategy and enhancing the adaptability to sudden noise environments.

[0049] In specific implementation, the trend prediction model of the noise disturbance of the speech signal in the hearing aid is trained based on the transient noise dynamic characteristics and the speech masking index. The following method is adopted, namely: first, a trend prediction model is initialized; second, the severity of the noise disturbance of the current speech in the hearing aid is evaluated according to the transient noise dynamic characteristics; the learning rate (LR) of the trend prediction model of the noise disturbance of the speech signal in the hearing aid is dynamically adjusted during the training process, so that the model can increase the parameter update speed in the drastic change area and reduce the update amplitude in the stable section to avoid oscillation, thereby improving the convergence efficiency and short-term prediction accuracy of the model; then, the degree of mixing of speech and noise in the ambient sound wave signal received by the hearing aid is evaluated according to the speech masking index, and the weight coefficients of the speech and noise error components in the loss function during the model training process are adjusted accordingly, so that the model pays more attention to the retention of speech information in the high coupling state and pays more attention to the modeling of the noise interference trend in the low coupling state, thereby improving the discrimination robustness and generalization ability of the model, so that the trend The prediction model more effectively adapts to the trend modeling requirements of voice quality changes under different noise interference backgrounds. In other embodiments, other methods can also be used for model training, which are not limited here; the disturbance trend prediction of the voice signal noise in the hearing aid through the trained trend prediction model can be achieved in the following way, namely: the ambient sound wave signal received by the hearing aid is input into the trained trend prediction model, thereby outputting the prediction result of the noise disturbance change trend of the voice signal in the hearing aid. The prediction result can provide feedforward information for the subsequent dynamic adjustment of the adaptive filter coefficient, thereby realizing active noise reduction control with early intervention and reduced delay.

[0050] In specific implementation, updating the adaptive filter coefficient of the hearing aid based on the prediction result can be achieved in the following way, namely: first, the prediction result output by the trend prediction model of the speech signal noise disturbance in the hearing aid is matched and analyzed with the adaptive filter coefficient currently being used by the hearing aid to determine whether the current coefficient meets the upcoming noise disturbance scenario; if the prediction result shows that the noise interference level will increase in the short term in the future, the current filter gain value is increased through the adaptive parameter adjustment mechanism, especially the frequency suppression weight is increased in the noise-dominated frequency band to enhance the shielding effect of high-intensity noise; if the prediction result shows that the noise disturbance tends to weaken in the future, the adaptive filter coefficient is reduced to reduce speech distortion, thereby improving speech intelligibility; wherein, the coefficient update process is implemented based on a preset dynamic adjustment function, the input of which is the trend prediction output value and the historical noise reduction performance index (such as signal-to-noise ratio, clarity index), and the output is the optimized adaptive filter coefficient; then the optimized adaptive filter coefficient is written back to the hearing aid digital signal processing (Digital Signal Processing) In the corresponding filtering algorithm control port in the SignalProcessor (DSP) module, it is ensured that the updated filtering coefficient is used for the next cycle signal processing, thereby realizing the adaptive dynamic noise reduction adjustment capability of the hearing aid; in other embodiments, other methods can also be used to improve the speech clarity of the hearing aid in the sudden noise environment and maintain the user's auditory comfort experience, which is not limited here.

[0051] In addition, in another aspect of the present application, in some embodiments, the present application provides a hearing aid intelligent protection device integrated on a head cap, the device including an adaptive noise reduction unit, Figure 4 , which is a schematic diagram of the structure of an adaptive noise reduction unit according to some embodiments of the present application. The adaptive noise reduction unit 400 includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described as follows: Acquisition module 401, in this application, the acquisition module 401 is mainly used to collect ambient sound wave signals through intelligent sensors integrated in the corresponding positions of the ears on both sides of the headwear, and transmit the collected ambient sound wave signals to the signal processor in the interlayer of the headwear; Processing module 402, in this application, the processing module 402 is mainly used to convert the ambient sound wave signal into a frequency domain signal by the signal processor, and then separate the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal; The processing module 402 in the present application is further configured to determine a speech masking index between the ambient noise and the speech in the ambient sound wave signal based on a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; The processing module 402 in the present application is further configured to determine the noise interference experienced by the hearing aid in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and determine the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the mutation amplitude of the noise in the noise interference; The execution module 403 in this application is mainly used to predict the disturbance trend of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index, and then update the adaptive filter coefficient of the hearing aid according to the prediction result.

[0052] The above describes in detail the examples of the intelligent protection device and method for hearing aids integrated on a headgear provided by the embodiments of the present application. It can be understood that in order to realize the above functions, the corresponding device includes hardware structures and / or software modules corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the units and algorithm steps of the various examples described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0053] In some embodiments, the present application also provides a computer device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device performs the above-mentioned hearing aid adaptive noise reduction method.

[0054] In some embodiments, reference Figure 5 The dotted line in the figure indicates that the unit or module is optional. The figure is a schematic diagram of the structure of a computer device that implements the hearing aid adaptive noise reduction method of the present application. The hearing aid adaptive noise reduction method in the above embodiment can be Figure 5 The computer device 500 is implemented as shown, and the computer device 500 includes at least one processor 501, a memory 502 and at least one communication unit 505. The computer device 500 can be a terminal device, a server or a chip.

[0055] The processor 501 may be a general-purpose processor or a special-purpose processor. For example, the processor 501 may be a central processing unit (CPU), which may be used to control the computer device 500, execute software programs, and process data from the software programs. The computer device 500 may also include a communication unit 505 for inputting (receiving) and outputting (transmitting) signals.

[0056] For example, the computer device 500 may be a chip, the communication unit 505 may be an input and / or output circuit of the chip, or the communication unit 505 may be a communication interface of the chip, and the chip may be a component of a terminal device, a network device, or other device.

[0057] For another example, the computer device 500 may be a terminal device or a server, and the communication unit 505 may be a transceiver of the terminal device or the server, or the communication unit 505 may be a transceiver circuit of the terminal device or the server.

[0058] The computer device 500 may include one or more memories 502, on which a program 504 is stored. The program 504 can be executed by the processor 501 to generate instructions 503, so that the processor 501 executes the method described in the above method embodiment according to the instructions 503. Optionally, data (such as a target audit model) can also be stored in the memory 502. Optionally, the processor 501 can also read data stored in the memory 502. The data can be stored at the same storage address as the program 504, or at a different storage address from the program 504.

[0059] The processor 501 and the memory 502 may be provided separately or integrated together, for example, integrated on a system on chip (SOC) of a terminal device.

[0060] It should be understood that each step of the above method embodiment can be completed by a hardware-based logic circuit or software-based instructions in the processor 501. The processor 501 can be a CPU, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, such as discrete gates, transistor logic devices, or discrete hardware components.

[0061] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0062] For example, in some embodiments, the present application also provides a computer-readable storage medium, which stores instructions or codes. When the instructions or codes are run on a computer, the computer implements the above-mentioned hearing aid adaptive noise reduction method.

[0063] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0064] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A hearing aid adaptive noise reduction method, which is used for an intelligent hearing aid protective device integrated on a headgear to perform adaptive noise reduction on the hearing aid, wherein: The hearing aid intelligent protection device integrated on the headgear includes an intelligent sensor and a signal processor, and is characterized in that the method includes the following steps: The intelligent sensors integrated in the corresponding positions of the ears on both sides of the head cap collect the environmental sound wave signals, and transmit the collected environmental sound wave signals to the signal processor in the interlayer of the head cap; The signal processor converts the ambient sound wave signal into a frequency domain signal, and then separates the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal; Determining a speech masking index between ambient noise and speech in the ambient sound wave signal based on a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; Determining the noise interference experienced by the hearing aid in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and determining the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the mutation amplitude of the noise in the noise interference; The disturbance trend of the speech signal noise in the hearing aid is predicted based on the transient noise dynamic characteristics and the speech masking index, and the adaptive filter coefficient of the hearing aid is updated according to the prediction result.

2. The method according to claim 1, wherein The signal processor converting the ambient sound wave signal into a frequency domain signal specifically includes: The signal processor divides the ambient sound wave signal into frames to obtain a plurality of audio frames; Perform frequency domain conversion on each audio frame to obtain a frequency domain signal of the ambient sound wave signal.

3. The method according to claim 1, wherein Separating the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise features and speech features in the frequency domain signal specifically includes: extracting noise features and speech features from the frequency domain signal; matching a noise-dominant frame in the hearing aid from the frequency domain signal according to the noise characteristics; A speech dominant frame in the hearing aid is matched from the frequency domain signal based on the speech feature.

4. The method according to claim 3, wherein Extracting noise features and speech features from the frequency domain signal specifically includes: Determining a spectral entropy value of each audio frame in the frequency domain signal; determining a speech band distribution of the frequency domain signal based on all spectral entropy values; Noise features and speech features are extracted from the frequency domain signal according to the speech band distribution.

5. The method according to claim 1, wherein Determining the speech masking index between the ambient noise and the speech in the ambient sound wave signal according to the distribution correlation relationship between the noise-dominated frame and the speech-dominated frame specifically includes: Determining a ratio distribution diagram of the noise-dominated frames and the speech-dominated frames; extracting a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames from the proportion distribution graph; Determining an inter-frame overlap between the noise-dominated frame and the speech-dominated frame based on the distribution association relationship; The speech masking index between the ambient noise and the speech in the ambient sound wave signal is determined according to the inter-frame overlap.

6. The method according to claim 1, wherein Determining the noise interference suffered by the hearing aid in the current wearing scenario according to the energy distribution characteristics of the noise spectrum in the frequency domain signal specifically includes: determining energy distribution characteristics of a noise spectrum in the frequency domain signal; determining a noise classification index of the frequency domain signal based on the energy distribution characteristics; The noise interference to which the hearing aid is subjected in the current wearing scenario is determined by the noise classification index.

7. The method according to claim 1, wherein Predicting the disturbance trend of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index specifically includes: training a trend prediction model of speech signal noise disturbance in a hearing aid based on the transient noise dynamic characteristics and the speech masking index; The disturbance trend of speech signal noise in the hearing aid is predicted using the trained trend prediction model.

8. An intelligent hearing aid protection device integrated into a headgear, the device including an adaptive noise reduction unit, characterized in that: The adaptive noise reduction unit includes: The acquisition module is used to collect ambient sound wave signals through intelligent sensors integrated in the corresponding positions of the ears on both sides of the headwear, and transmit the collected ambient sound wave signals to the signal processor in the interlayer of the headwear; a processing module, configured to convert the ambient sound wave signal into a frequency domain signal by the signal processor, and then separate the noise-dominated frame and the speech-dominated frame in the hearing aid based on the noise characteristics and speech characteristics in the frequency domain signal; The processing module is further configured to determine a speech masking index between the ambient noise and the speech in the ambient sound wave signal based on a distribution correlation relationship between the noise-dominated frames and the speech-dominated frames; The processing module is further configured to determine the noise interference experienced by the hearing aid in the current wearing scenario based on the energy distribution characteristics of the noise spectrum in the frequency domain signal, and determine the transient noise dynamic characteristics of the ambient sound wave signal under the current noise interference based on the mutation amplitude of the noise in the noise interference; An execution module is used to predict the disturbance trend of the speech signal noise in the hearing aid based on the transient noise dynamic characteristics and the speech masking index, and then update the adaptive filter coefficient of the hearing aid according to the prediction result.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the hearing aid adaptive noise reduction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions or codes, and when the instructions or codes are executed on a computer, the computer implements the hearing aid adaptive noise reduction method according to any one of claims 1 to 7.

Images