CN-121985278-A - Hearing aid auxiliary system based on artificial intelligence

CN121985278ACN 121985278 ACN121985278 ACN 121985278ACN-121985278-A

Abstract

The invention discloses an artificial intelligence-based hearing aid auxiliary system, which belongs to the technical field of hearing aid machinery and comprises a bone conduction hearing aid, a pickup module and a personal terminal, wherein the pickup module is arranged on the bone conduction hearing aid and is used for collecting surrounding environment sounds and extracting clear voice information in the surrounding environment sounds, and the personal terminal is used for controlling the bone conduction hearing aid to perform auxiliary hearing test and is used for controlling the pickup module. The hearing aid auxiliary system combines the technologies of voice processing, artificial intelligence and the like, can realize intelligent analysis and adaptation to the hearing demand and environment of a user, improves the personalized adaptation capability and effect of the hearing aid, and enables the user to obtain better hearing experience.

Inventors

YU LE
WANG MIN
LIANG JIXIANG
WANG BOYUAN
WANG XIN
LIU XUEMEI
REN YIFEI
GAO KAI
ZHAO YUTING
KANG YAO

Assignees

北京城市学院

Dates

Publication Date: 20260505
Application Date: 20260205

Claims (10)

1. The hearing aid auxiliary system based on artificial intelligence is characterized by comprising a bone conduction hearing aid, a pickup module and a personal terminal, wherein the pickup module is arranged on the bone conduction hearing aid and is used for collecting surrounding environment sounds and extracting clear voice information in the surrounding environment sounds; The sound pickup module comprises a sound pickup device, a noise reduction unit, a voice extraction unit, an information enhancement unit and a direction judgment unit, wherein the sound pickup device is used for collecting surrounding environment sounds, the noise reduction unit is used for carrying out noise reduction treatment on the surrounding environment sounds to obtain a processed sound signal, the voice extraction unit is used for extracting voice information from the processed sound signal, the information enhancement unit is used for carrying out information enhancement on the voice information to obtain clear voice information with a signal-to-noise ratio improvement of more than or equal to 15dB and a 5-minute voice intelligibility score of more than or equal to 3.5, and transmitting the clear voice information to the personal terminal, and the direction judgment unit is used for judging the sound emitting direction based on a self-adaptive wave beam method; The personal terminal comprises an environmental sound detection unit, a hearing test unit, a regulation and control unit, a conversion unit and an interaction unit, wherein the environmental sound detection unit is used for receiving surrounding environmental sounds and judging whether the current environment accords with the test environment, the hearing test unit is used for conducting hearing tests on a wearer and generating a hearing loss grade report, the hearing loss grade report is a hearing threshold condition report of a user, the regulation and control unit is used for generating a compensation scheme based on the hearing loss grade report and conducting sound compensation on the bone conduction hearing aid based on the compensation scheme and the sound emitting direction, the regulation and control unit is further used for controlling the pick-up direction of the pick-up module, the conversion unit is used for receiving clear voice information and converting the clear voice information into text information, the interaction unit is used for providing man-machine interaction, and the interaction unit is further used for displaying the text information.
2. The hearing assistance system based on artificial intelligence according to claim 1, wherein in the direction judging unit, by collecting a plurality of sound signals in a microphone array of the pickup device, a delay difference and a relative intensity difference of the signals are calculated, then the received plurality of sound signals are dynamically weighted by using a variable step-size normalized least mean square algorithm, a signal-to-wave noise ratio of the signals is improved, and then the weighted signals are combined to estimate a sound emitting direction.
3. The hearing assistance system based on artificial intelligence of claim 2, wherein the human voice extraction unit comprises a receiving device, a dereverberation device and a second DSP chip, wherein the receiving device is used for receiving the processed sound signal, the dereverberation device is used for dereverberating the processed sound signal to obtain a dereverberated sound signal, and the second DSP chip is used for blind source separation of the dereverberated sound signal to obtain the voice information.
4. An artificial intelligence based hearing aid system according to claim 3, wherein in the dereverberation apparatus, the dereverberation process is a short time fourier transform of the processed sound signal, converting the time domain signal into a frequency domain signal, and then applying a normalized least mean square adaptive algorithm to perform multi-channel dereverberation, effectively eliminating the environmental echo and reverberation effects, resulting in a dereverberated sound signal.
5. The hearing aid system based on artificial intelligence according to claim 3, wherein in the second DSP chip, the blind source separation is specifically performed by decomposing the dereverberated sound signal after dereverberated processing based on OverIVA blind source separation algorithm, separating out the target voice component by maximizing the independence between the signal sources and suppressing the background noise, thereby obtaining the purified target voice signal, and obtaining the voice information.
6. The hearing aid system based on artificial intelligence according to claim 3, wherein the noise reduction unit comprises a first low-pass filter device, a high-pass filter device, a first DSP chip and a second low-pass filter device, wherein the first low-pass filter device is used for filtering high-frequency noise in surrounding environment sound and inputting high-frequency phase signals into the first DSP chip, the high-pass filter device is used for filtering low-frequency noise in the surrounding environment sound and inputting low-frequency phase signals into the first DSP chip, the first DSP chip is used for carrying out phase superposition on the high-frequency phase signals and the low-frequency phase signals to obtain a superposed signal, and the second low-pass filter device is used for filtering high-frequency residual noise of the superposed signal to obtain the processed sound signal.
7. The hearing assistance system based on artificial intelligence according to claim 6, wherein the information enhancement unit enhances the voice information based on a least squares GAN model, specifically by denoising and spectrum compensating the input voice information using a generator G, discriminating the generated result based on a least squares loss function by a discriminator D, performing countermeasure training by the generator and the discriminator, and finally outputting clear voice information by the generator G.
8. The hearing assistance system based on artificial intelligence according to claim 7, wherein the adjusting unit dynamically adjusts the amplification curve of the bone conduction hearing aid based on the frequency adaptive gain model generated by the hearing loss level report to obtain a compensation scheme by: S21, extracting hearing threshold values of each core frequency in the frequency range of 125Hz-8kHz from a hearing loss grade report Average hearing threshold ; S22, calculating target gains of all frequency bands based on NAL-NL2 prescription formula Constructing a frequency self-adaptive gain model; S23, combining the sound source angle output by the direction judging unit Performing spatial filtering optimization through MVDR beam former weight vectors; S24, converting target gain and weight vectors of each frequency band into amplification curve adjustment parameters of the bone conduction hearing aid, adapting to a built-in amplification channel of the hearing aid, setting a maximum gain safety threshold value, and updating an amplification curve; And S25, integrating the adjusted amplification curve parameters into a complete compensation scheme, and transmitting the complete compensation scheme to an audio output module of the bone conduction hearing aid.
9. The artificial intelligence based hearing assistance system of claim 8 wherein the frequency adaptive gain model calculates target gains for each frequency band based on NAL-NL2 prescription formulas , Wherein For frequency Is used for the hearing threshold value of (a), Is the average hearing threshold.
10. An artificial intelligence based hearing assistance system according to claim 8, wherein in the hearing test unit the hearing test procedure is as follows: generating a plurality of test audios with different frequencies and decibels, and transmitting the test audios to a user through the bone conduction hearing aid; The user selects whether the test audio can be heard or not according to the prompt of the interaction unit; and calculating a hearing threshold value at each frequency by adopting a cubic spline interpolation method according to the selection condition of a user, and generating the hearing loss grade report.

Description

Hearing aid auxiliary system based on artificial intelligence Technical Field The invention relates to the technical field of hearing aids, in particular to an artificial intelligence-based hearing aid auxiliary system. Background Hearing loss can seriously affect the physical and mental health of hearing impaired patients, and wearing a hearing aid is the most effective means for improving hearing of hearing impaired patients at present. The existing hearing aid technology plays an important role in improving the hearing effect and the quality of life of the hearing impaired people. However, due to the variety of different hearing demands and living environments that exist for hearing impaired individuals, conventional hearing aid technology has difficulty meeting their individual needs and challenges to complex environments. Therefore, the invention provides an artificial intelligence-based hearing aid auxiliary system to improve the adaptability, intelligence and user experience of a hearing aid. Disclosure of Invention The invention aims to solve the technical problems of improving the adaptability, the intelligence and the user experience of a hearing aid, providing an artificial intelligence-based hearing aid auxiliary system, processing acquired audio information by utilizing OverIVA algorithm and least square GAN model to obtain clear voice information, simultaneously carrying out hearing test on a user, and formulating a sound compensation scheme based on the test result. The invention solves the technical problems through the following technical scheme that the bone conduction hearing aid comprises a bone conduction hearing aid, a pickup module and a personal terminal, wherein the pickup module is arranged on the bone conduction hearing aid and is used for collecting surrounding environment sounds and extracting clear voice information in the surrounding environment sounds; The sound pickup module comprises a sound pickup device, a noise reduction unit, a voice extraction unit, an information enhancement unit and a direction judgment unit, wherein the sound pickup device is used for collecting surrounding environment sounds, the noise reduction unit is used for carrying out noise reduction treatment on the surrounding environment sounds to obtain a processed sound signal, the voice extraction unit is used for extracting voice information from the processed sound signal, the information enhancement unit is used for carrying out information enhancement on the voice information to obtain clear voice information with a signal-to-noise ratio improvement of more than or equal to 15dB and a 5-minute voice intelligibility score of more than or equal to 3.5, and transmitting the clear voice information to the personal terminal, and the direction judgment unit is used for judging the sound emitting direction based on a self-adaptive wave beam method; The personal terminal comprises an environmental sound detection unit, a hearing test unit, a regulation and control unit, a conversion unit and an interaction unit, wherein the environmental sound detection unit is used for receiving surrounding environmental sounds and judging whether the current environment accords with the test environment, the hearing test unit is used for conducting hearing tests on a wearer and generating a hearing loss grade report, the hearing loss grade report is a hearing threshold condition report of a user, the regulation and control unit is used for generating a compensation scheme based on the hearing loss grade report and conducting sound compensation on the bone conduction hearing aid based on the compensation scheme and the sound emitting direction, the regulation and control unit is further used for controlling the pick-up direction of the pick-up module, the conversion unit is used for receiving clear voice information and converting the clear voice information into text information, the interaction unit is used for providing man-machine interaction, and the interaction unit is further used for displaying the text information. Furthermore, in the direction judging unit, the time delay difference and the relative intensity difference of the signals are calculated by collecting a plurality of sound signals in a microphone array of the pickup device, then the received sound signals are dynamically weighted by utilizing a variable step length normalized least mean square algorithm, the signal wave front signal to noise ratio is improved, and then the weighted signals are combined to estimate the sound emitting direction. Still further, the human voice extraction unit comprises a receiving device, a dereverberation device and a second DSP chip, wherein the receiving device is used for receiving the processed sound signals, the dereverberation device is used for carrying out dereverberation processing on the processed sound signals to obtain dereverberation sound signals, and the second DSP chip is used for carrying out blind sour