CN106714062A - BP-artificial-neural-network-based intelligent matching algorithm for digital hearing aid - Google Patents
BP-artificial-neural-network-based intelligent matching algorithm for digital hearing aid Download PDFInfo
- Publication number
- CN106714062A CN106714062A CN201611087426.0A CN201611087426A CN106714062A CN 106714062 A CN106714062 A CN 106714062A CN 201611087426 A CN201611087426 A CN 201611087426A CN 106714062 A CN106714062 A CN 106714062A
- Authority
- CN
- China
- Prior art keywords
- artificial neural
- neural networks
- algorithm
- network
- gain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R25/00—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception
- H04R25/50—Customised settings for obtaining desired overall acoustical characteristics
- H04R25/505—Customised settings for obtaining desired overall acoustical characteristics using digital signal processing
- H04R25/507—Customised settings for obtaining desired overall acoustical characteristics using digital signal processing implemented by neural network or fuzzy logic
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R2225/00—Details of deaf aids covered by H04R25/00, not provided for in any of its subgroups
- H04R2225/43—Signal processing in hearing aids to enhance the speech intelligibility
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Automation & Control Theory (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Neurosurgery (AREA)
- Otolaryngology (AREA)
- Acoustics & Sound (AREA)
- Signal Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a BP-artificial-neural-network-based intelligent matching algorithm for a digital hearing aid. On the basis of a BP artificial neural network, a network is trained by using lots of training data to obtain a satisfactory mature network; and the network is corrected by using a self-built formula model, thereby obtaining a mature intelligent matching algorithm. On the basis of the genetic algorithm principle, an initial weight value and a threshold of a BP artificial neural network are optimized; the BP neural network is trained by using an existing audiogram and a spectrum gain response as raining data; and the network is corrected based on a matching formula model to obtain a mature BP artificial neural network to replace the existing matching prescription formula, so that parameters like all channel gains, the maximum sound output, the compression rate, and the compression inflection point of the digital hearing aid can be obtained.
Description
Technical field
It is to be related to one kind manually refreshing based on BP in particular the invention belongs to testing with algorithm for digital deaf-aid field
Intelligently tested with algorithm through the digital deaf-aid of network, the algorithm is based on BP artificial neural networks, by substantial amounts of training data pair
Network is trained, and draws a gratifying maturation network, and network is repaiied using the formula model voluntarily set up
Just, tested with algorithm so as to obtain the intelligence of maturation.
Background technology
At present, Chinese population Aging Problem is on the rise, will be close to 30%, ear to the year two thousand thirty Aged in China population ratio
Deaf is common complaint among the elderly, and with the increase of elderly population, the quantity with deaf elderly population is increasing, people couple
It is growing day by day in the demand of audiphone.
The birth of digital deaf-aid has established solid foundation for the development of audiphone, either the standard of Speech processing
The convenience that true property is still applied has significant progress, and the optimization of these performances depends primarily on dynamic range compressor,
And dynamic compressor is the gain parameter depended primarily on obtained by testing with formula how to work.Public affairs are matched somebody with somebody in testing for digital deaf-aid
Formula has a lot, be broadly divided into it is based on the threshold of audibility and based on loudness, wherein being divided into based on the threshold of audibility is linear again
And it is nonlinear two kinds.For mainly having LGOB based on loudness, linear mainly has NAL, DSL etc., nonlinear
Mainly include NAL-NL1, FIG6, DSL (i/o) etc., most widely used at present be POGO, NAL, DSL these three.
Different testing is differed with formula for the result that same patient is drawn, tests also each with the effect for reaching
There are pros and cons.POGO formula are simple half gain methods, that is, the half for obtaining patient's threshold of audibility is obtained plus empirical
Channel gain, it is this test to be mainly with formula loudness is adjusted, the requirement for the intelligibility of language is less.NAL's
Four generations of development experience, present NAL-NL2 is highly developed, and the apolegamy particularly with moderate deafness patient has
Important practical value.DSL is tested and is primarily adapted for use in children with formula.Though this several conventional testing respectively has advantage with formula,
The deafness patient effect serious for hearing impairment is poor, and because different patients needs to select different testing with prescription public affairs
Formula is tested with effect with what is be optimal, and this gives to test and brings great inconvenience with work.
Dispenser wish one kind simply test with scheme, can according to the audiogram of patient draw one it is very accurate
Test with parameter, and gradual perfection can be increased with case with testing, satisfactorily tested with effect so as to reach one.
The content of the invention
The invention aims to overcome deficiency of the prior art, there is provided a kind of number based on BP artificial neural networks
Word audiphone is intelligently tested with algorithm, and the algorithm is based on BP artificial neural networks, and network is instructed by substantial amounts of training data
Practice, draw a gratifying maturation network, and network is modified using the formula model voluntarily set up, so as to obtain
Ripe intelligence is tested with algorithm；The present invention is carried out using principle of genetic algorithm to the initial weight and threshold value of BP artificial neural networks
Optimization, and BP artificial neural networks are trained as training data by the use of existing audiogram and spectral gain response, together
When network is modified with formula model by testing, the BP artificial neural networks of maturation are obtained to be tested with prescription instead of existing
Formula, and then obtain the parameters such as each channel gain, maximum voice output, compression ratio and the compression flex point of digital deaf-aid.
The purpose of the present invention is achieved through the following technical solutions：
A kind of digital deaf-aid based on BP artificial neural networks is intelligently tested with algorithm, using existing audiogram and frequency spectrum
Gain response goes BP ANN as training data, and the new patient of prediction is estimated by the network after training,
Specifically include following steps：
(1) BP artificial neural networks, including input layer, hidden layer and output layer are built；
(2) initial weight and initial threshold of the BP artificial neural networks set up to step (1) using genetic algorithm are carried out
Optimization；
(3) to being trained by the BP artificial neural networks after optimization in step (2), training data uses actual patient
Case, by the continuous training to BP artificial neural networks, make it progressively ripe, finally give relative maturity based on
The digital deaf-aid of BP artificial neural networks is tested with algorithm；
(4) build for by testing the formula model that the gain output obtained with algorithm is optimized step (3) Suo Shu；
(5) formula model set up using step (4) is modified with algorithm to described testing by weighting scheme, with
Gain output is obtained, with being on the increase for follow-up training data, weight coefficient can move closer to 1, finally given ripe perfect
The digital deaf-aid based on BP artificial neural networks intelligently test with algorithm.
BP artificial neural networks parameter described in step (1) includes the weights W between input layer and hidden layerijAnd it is implicit
Weights W between layer and output layerjk, the threshold value also including hidden layer and output layer.
The nodes of the input layer, hidden layer and output layer are respectively 10,20 and 30, and being divided into 10 according to frequency spectrum leads to
Road, i.e. frequency are the increasing of 40dB in the target of 250,500,750,1000,1500,2000,3000,4000,6000 and 8000Hz
Benefit, compression ratio and compression flex point.
Optimizing Flow is as follows described in step (2)：A. the population scale and iterations of genetic algorithm are set；B. by BP people
The initial weight and threshold value of artificial neural networks carry out GA codings, are obtained per each and every one by audiogram and the spectral gain response of patient
The fitness of body；C. selection operation, crossover operation and the mutation operation for eventually passing through genetic algorithm enter to initial weight and threshold value
The calculating of row population's fitness, is finally reached the optimizing to network threshold and weights.
Compared with prior art, the beneficial effect that technical scheme is brought is：
1., in order to the initial weight and threshold value of neutral net is determined more accurately, the present invention is using genetic algorithm to nerve
Network is optimized, and the optimized algorithm is highly developed, and the principle that genetic algorithm follows " number competing people select, the superior's existence " is completed
Optimizing to network initial weight and threshold value；It is artificial relative to original BP after BP artificial neural networks are through genetic algorithm optimization
Neutral net possesses inborn advantage, and network is carried out with the audiogram and spectral gain response in case by will largely test
Training, with being continuously increased for training data, the digital deaf-aid based on BP artificial neural networks is intelligently tested also will more with algorithm
Come more ripe.
2., in order that must test that the result matched somebody with somebody is more accurate, the present invention is also using the formula model oneself set up to the artificial god of BP
It is corrected through network, each channel gain is drawn by both weightings, finally gives a digital deaf-aid intelligence for maturation
Can test with algorithm.
3. testing with the stage, dispenser only needs to as input data test the result of hearing test through intelligence of the invention and matches somebody with somebody
Algorithm is calculated just can draw the parameters such as gratifying each channel gain, compression ratio and compression flex point, because being based on BP people
The digital deaf-aid algorithm of artificial neural networks is that the result of its output must by largely testing the ripe network obtained with case training
So can really be tested with effect close to patient.
4. the algorithm model that the present invention relates to relative maturity, tests with prescriptive formula compared to others, is based on
It is more extensive that the digital deaf-aid of BP artificial neural networks intelligently tests the scope covered with algorithm, and due to the algorithm model
It is that, with the ripe algorithm model obtained by case, the result of its output also will to a certain extent be better than other by true experiment
Test with prescriptive formula.The algorithm that the present invention relates to be suitable for digital deaf-aid test match somebody with somebody during replace existing testing with prescription
Formula, will provide one and more accurately easily test with scheme to test with work.
Brief description of the drawings
Fig. 1 is the topological structure schematic diagram of the BP artificial neural networks used in the present invention.
Fig. 2 is that the present invention utilizes genetic algorithm optimization BP artificial neural network and the particular flow sheet being trained.
Fig. 3 is that the formula model used in the present invention is tested with scheme flow chart.
Fig. 4 is that the formula used in the present invention tests the flow chart tested with principle and the weighting of BP neural network intelligent algorithm and matched somebody with somebody.
Specific embodiment
The invention will be further described below in conjunction with the accompanying drawings：
A kind of digital deaf-aid based on BP artificial neural networks is intelligently tested with algorithm, actually using existing hearing
Figure and spectral gain response go BP ANN as training data, estimate that prediction is new by the network after training
Patient.It is exactly further using the BP artificial neural networks optimized by principle of genetic algorithm, by existing audiogram
BP artificial neural networks are trained as training data with spectral gain response, at the same using test the model with formula to instruction
Network after white silk is modified, and the digital deaf-aid based on BP artificial neural networks for finally giving maturation is tested with algorithm.Specifically
Process steps are as follows：
(1) build BP artificial neural networks, BP artificial neural networks that the present embodiment is used it is a kind of it is multi-level before
To transmission before feedback neutral net, i.e. data, to transmission after error.As shown in figure 1, be the topological structure of BP artificial neural networks,
BP artificial neural networks are a kind of highly developed artificial neural networks, and primary structure includes input layer, hidden layer and defeated
Go out layer.Major parameter includes the weights W between input layer and hidden layerijAnd the weights W between hidden layer and output layerjk, together
When also include the threshold value of hidden layer and output layer.The input section for specifying BP artificial neural networks with algorithm model is tested in the present invention
Point for the threshold of audibility in 250,500,1000,2000,3000,4000,6000,8000Hz of 10, i.e. frequency and sex (with 1,
2 represent man and female respectively) and experience (represent user, Short-term user, expert user and long-term use first respectively with 1,2,3,4
Family).Node in hidden layer is 20, and output node number is 30, be output as the output of 10 passages, i.e. frequency 250,500,750,
1000th, the target of 1500,2000,3000,4000,6000 and 8000Hz is gain, compression ratio and the compression flex point of 40dB.For
The training data of this network is trained to be all from the true experiment of patient with case, it is ensured that the reliability of network.
(2) initial weight and initial threshold of the BP artificial neural networks set up to step (1) are optimized；Fig. 2 be into
Row optimization and the particular flow sheet that is trained to network, have been specifically described the BP that this algorithm utilized manually refreshing in Fig. 1
Through the design parameter of network, in order to ensure the optimality of network, the present invention is carried out by genetic algorithm to BP artificial neural networks
Optimization.Genetic algorithm is the algorithm that a kind of logarithm value is optimized, herein it is intended that neutral net seeks optimal initial weight
And threshold value.The population scale for specifying genetic algorithm in the present invention is 50 individualities, and iterations is 1000 times.Specific Optimization Steps
It is that the initial weight and threshold value of BP artificial neural networks are carried out into GA codings, then by testing the audiogram with the patient in case
Each individual fitness is obtained with spectral gain response, then by the selection operation of genetic algorithm, crossover operation and variation
Operation carries out the calculating of population's fitness to initial weight and threshold value, is finally reached the optimizing to network threshold and weights.In instruction
During white silk, BP neural network passes sequentially through the calculating of hidden layer node and the meter of output node according to the audiometric data of input
Calculation draws i.e. each channel gain of prediction output, compression ratio and compression flex point, and prediction output is compared by with desired output
Go out error, error carries out threshold value and weights in back transfer renewal network by network so that the result of output is gradually approached
Real output result.
(3) to being trained by the BP artificial neural networks after optimization in step (2), training data uses actual patient
Case, by the continuous training to BP artificial neural networks, make its step ripe, finally give relative maturity based on BP
The digital deaf-aid of artificial neural network is tested with algorithm；
(4) build for by testing the formula model that the gain output obtained with algorithm is optimized step (3) Suo Shu；
Fig. 3 is that the formula model built tests the flow chart with scheme.Input data is obtained first, mainly there are the eight i.e. frequencies of the passage threshold of audibility to exist
250th, 500 the threshold of audibility, 1000,2000,3000,4000,6000, during 8000Hz and whether be conduction deafness patient (it is 1 to be,
It is not gain 0), then to take when the half of the threshold of audibility is input into as 40dB and ten passages will be changed to using mean value method
(frequency is 250,500,750,1000,1500,2000,3000,4000,6000, the gain of 8000Hz), will frequency be 500
Averaged the gain as 750Hz with the gain of 1000Hz, the gain at 1000 and 2000Hz is taken average as 1500Hz
Gain, while according to whether for conduction deafness patient carry out plus 6dB modification, finally by compression ratio (conduction deafness suffer from
Person takes 1.2, and gain when 1.4) phonosensitive nerve deafness patient takes and compression flex point is input into 60dB and 80dB is calculated, together
When to input 80dB when 250Hz places gain carry out reduction 6dB, gain near 3000~4000Hz reduction 3dB, while root
According to input threshold of audibility estimation MPO.
(5) formula model set up using step (4) is modified with algorithm to described testing by weighting scheme, is schemed
4 is that the formula that the present invention is used tests the flow chart for being weighted with principle and the intelligent algorithm of BP artificial neural networks and testing and match somebody with somebody, plus
The purpose of power is modified in order that obtaining the output result of intelligent algorithm based on BP artificial neural networks so that output it is each
Channel gain more approaching to reality level.Specific method be first according to input obtain formula test with result and intelligence test with algorithm
Test with result, at the same the result to generating carry out treatment generation 3X10 gain matrix (frequency is 250,500,750,1000,
1500th, 2000,3000,4000,6000, during 8000Hz input be 40,60,80dB when gain), then formula is tested with obtaining
Matrix and the matrix that obtains of intelligent algorithm in each gain be weighted and obtain total gain matrix, finally according to MPO
Result to exporting is limited, and wherein weight coefficient p is falling for a matrix of the 1X10 related to network, i.e. neutral net
Number is multiplied by a constant k, as the gradually maturation k values of network can become larger, until all coefficients in p are changed into 1, now
Indicate the maturation of network, digital deaf-aid of the present invention based on BP artificial neural networks intelligently tests just ripe and perfect with algorithm
.
The present invention is not limited to embodiments described above.Description to specific embodiment above is intended to describe and says
Bright technical scheme, above-mentioned specific embodiment is only schematical, is not restricted.This is not being departed from
In the case of invention objective and scope of the claimed protection, one of ordinary skill in the art may be used also under enlightenment of the invention
The specific conversion of many forms is made, these are belonged within protection scope of the present invention.
Claims (4)
1. a kind of digital deaf-aid based on BP artificial neural networks is intelligently tested with algorithm, it is characterised in that listened using existing
Try hard to go BP ANN as training data with spectral gain response, prediction is estimated by the network after training
New patient, specifically includes following steps：
(1) BP artificial neural networks, including input layer, hidden layer and output layer are built；
(2) initial weight and initial threshold of the BP artificial neural networks set up to step (1) using genetic algorithm are optimized；
(3) to being trained by the BP artificial neural networks after optimization in step (2), training data uses the case of actual patient
Example, by the continuous training to BP artificial neural networks, makes it progressively ripe, finally give relative maturity based on BP people
The digital deaf-aid of artificial neural networks is tested with algorithm；
(4) build for by testing the formula model that the gain output obtained with algorithm is optimized step (3) Suo Shu；
(5) formula model set up using step (4) is modified with algorithm to described testing by weighting scheme, to obtain
Gain is exported, and with being on the increase for follow-up training data, weight coefficient can move closer to 1, finally gives ripe perfect base
Intelligently tested with algorithm in the digital deaf-aid of BP artificial neural networks.
2. a kind of digital deaf-aid based on BP artificial neural networks is intelligently tested with algorithm according to claim 1, its feature
It is that BP artificial neural networks parameter described in step (1) includes the weights W between input layer and hidden layerijAnd hidden layer
Weights W and output layer betweenjk, the threshold value also including hidden layer and output layer.
3. a kind of digital deaf-aid based on BP artificial neural networks according to claim 1 or claim 2 is intelligently tested with algorithm, and it is special
Levy and be, the nodes of the input layer, hidden layer and output layer are respectively 10,20 and 30,10 passages are divided into according to frequency spectrum,
I.e. frequency 250,500,750,1000,1500,2000,3000,4000,6000 and 8000Hz target for 40dB gain,
Compression ratio and compression flex point.
4. a kind of digital deaf-aid based on BP artificial neural networks is intelligently tested with algorithm according to claim 1, its feature
It is that Optimizing Flow is as follows described in step (2)：
A. the population scale and iterations of genetic algorithm are set；
B. the initial weight and threshold value of BP artificial neural networks are carried out into GA codings, is rung by the audiogram and spectrum gain of patient
Each individual fitness should be obtained；
C. eventually pass through selection operation, crossover operation and the mutation operation of genetic algorithm initial weight and threshold value are carried out population fit
The calculating of response, is finally reached the optimizing to network threshold and weights.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611087426.0A CN106714062B (en) | 2016-11-30 | 2016-11-30 | Digital hearing aid intelligent fitting method based on BP artificial neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611087426.0A CN106714062B (en) | 2016-11-30 | 2016-11-30 | Digital hearing aid intelligent fitting method based on BP artificial neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106714062A true CN106714062A (en) | 2017-05-24 |
| CN106714062B CN106714062B (en) | 2020-02-18 |
Family
ID=58935394
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611087426.0A Active CN106714062B (en) | 2016-11-30 | 2016-11-30 | Digital hearing aid intelligent fitting method based on BP artificial neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106714062B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108703761A (en) * | 2018-06-11 | 2018-10-26 | 佛山博智医疗科技有限公司 | The test method of Auditory identification susceptibility |
| CN109147808A (en) * | 2018-07-13 | 2019-01-04 | 南京工程学院 | A kind of Speech enhancement hearing-aid method |
| CN109151692A (en) * | 2018-07-13 | 2019-01-04 | 南京工程学院 | Hearing aid based on deep learning network tests method of completing the square certainly |
| CN109714692A (en) * | 2018-12-26 | 2019-05-03 | 天津大学 | Noise reduction method based on personal data and artificial neural network |
| CN110166917A (en) * | 2018-02-16 | 2019-08-23 | 西万拓私人有限公司 | Method for adjusting the parameter of hearing system |
| CN110473567A (en) * | 2019-09-06 | 2019-11-19 | 上海又为智能科技有限公司 | Audio-frequency processing method, device and storage medium based on deep neural network |
| CN111491245A (en) * | 2020-03-13 | 2020-08-04 | 天津大学 | Digital hearing aid sound field identification algorithm based on cyclic neural network and hardware implementation method |
| CN111818436A (en) * | 2020-07-14 | 2020-10-23 | 无锡清耳话声科技有限公司 | Real ear analysis test system based on machine learning |
| CN112383870A (en) * | 2020-10-29 | 2021-02-19 | 惠州市锦好医疗科技股份有限公司 | Adaptive hearing parameter fitting method and device |
| CN112887885A (en) * | 2021-01-12 | 2021-06-01 | 天津大学 | Hearing aid fault automatic detection system and hearing aid system |
| CN116614757A (en) * | 2023-07-18 | 2023-08-18 | 江西斐耳科技有限公司 | Hearing aid fitting method and system based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1523219B1 (en) * | 2003-10-10 | 2008-08-20 | Siemens Audiologische Technik GmbH | Method for training and operating a hearing aid and corresponding hearing aid |
| CN104053112A (en) * | 2014-06-26 | 2014-09-17 | 南京工程学院 | Hearing aid self-fitting method |
| CN105611477A (en) * | 2015-12-27 | 2016-05-25 | 北京工业大学 | Depth and breadth neural network combined speech enhancement algorithm of digital hearing aid |
| CN105722001A (en) * | 2014-12-23 | 2016-06-29 | 奥迪康有限公司 | Hearing Device Adapted For Estimating A Current Real Ear To Coupler Difference |
-
2016
- 2016-11-30 CN CN201611087426.0A patent/CN106714062B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1523219B1 (en) * | 2003-10-10 | 2008-08-20 | Siemens Audiologische Technik GmbH | Method for training and operating a hearing aid and corresponding hearing aid |
| CN104053112A (en) * | 2014-06-26 | 2014-09-17 | 南京工程学院 | Hearing aid self-fitting method |
| CN105722001A (en) * | 2014-12-23 | 2016-06-29 | 奥迪康有限公司 | Hearing Device Adapted For Estimating A Current Real Ear To Coupler Difference |
| CN105611477A (en) * | 2015-12-27 | 2016-05-25 | 北京工业大学 | Depth and breadth neural network combined speech enhancement algorithm of digital hearing aid |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110166917A (en) * | 2018-02-16 | 2019-08-23 | 西万拓私人有限公司 | Method for adjusting the parameter of hearing system |
| CN108703761A (en) * | 2018-06-11 | 2018-10-26 | 佛山博智医疗科技有限公司 | The test method of Auditory identification susceptibility |
| CN109147808A (en) * | 2018-07-13 | 2019-01-04 | 南京工程学院 | A kind of Speech enhancement hearing-aid method |
| CN109151692A (en) * | 2018-07-13 | 2019-01-04 | 南京工程学院 | Hearing aid based on deep learning network tests method of completing the square certainly |
| CN109147808B (en) * | 2018-07-13 | 2022-10-21 | 南京工程学院 | Speech enhancement hearing aid method |
| CN109714692A (en) * | 2018-12-26 | 2019-05-03 | 天津大学 | Noise reduction method based on personal data and artificial neural network |
| CN110473567A (en) * | 2019-09-06 | 2019-11-19 | 上海又为智能科技有限公司 | Audio-frequency processing method, device and storage medium based on deep neural network |
| CN111491245B (en) * | 2020-03-13 | 2022-03-04 | 天津大学 | Digital hearing aid sound field identification algorithm based on cyclic neural network and implementation method |
| CN111491245A (en) * | 2020-03-13 | 2020-08-04 | 天津大学 | Digital hearing aid sound field identification algorithm based on cyclic neural network and hardware implementation method |
| CN111818436A (en) * | 2020-07-14 | 2020-10-23 | 无锡清耳话声科技有限公司 | Real ear analysis test system based on machine learning |
| CN111818436B (en) * | 2020-07-14 | 2021-09-28 | 无锡清耳话声科技有限公司 | Real ear analysis test system based on machine learning |
| CN112383870A (en) * | 2020-10-29 | 2021-02-19 | 惠州市锦好医疗科技股份有限公司 | Adaptive hearing parameter fitting method and device |
| CN112383870B (en) * | 2020-10-29 | 2022-03-18 | 惠州市锦好医疗科技股份有限公司 | Adaptive hearing parameter fitting method and device |
| CN112887885A (en) * | 2021-01-12 | 2021-06-01 | 天津大学 | Hearing aid fault automatic detection system and hearing aid system |
| CN116614757A (en) * | 2023-07-18 | 2023-08-18 | 江西斐耳科技有限公司 | Hearing aid fitting method and system based on deep learning |
| CN116614757B (en) * | 2023-07-18 | 2023-09-26 | 江西斐耳科技有限公司 | Hearing aid fitting method and system based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106714062B (en) | 2020-02-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106714062A (en) | BP-artificial-neural-network-based intelligent matching algorithm for digital hearing aid | |
| CN107358966B (en) | No-reference speech quality objective assessment method based on deep learning speech enhancement | |
| US5729658A (en) | Evaluating intelligibility of speech reproduction and transmission across multiple listening conditions | |
| US6879860B2 (en) | Cochlear implant MAP optimization with use of a genetic algorithm | |
| CN103079160B (en) | One is automatically tested and is joined digital hearing-aid and method thereof | |
| CN105163262B (en) | A kind of loudspeaker sound detection method and detecting system | |
| Jorgensen | Verification and validation of hearing aids: Opportunity not an obstacle | |
| CN109330846B (en) | Air wave pressure massage instrument parameter optimization method based on deep learning algorithm | |
| CN111968677B (en) | Voice quality self-evaluation method for fitting-free hearing aid | |
| CN111696575B (en) | Low ventilation and apnea detection and identification system based on hybrid neural network model | |
| CN109714692A (en) | Noise reduction method based on personal data and artificial neural network | |
| Boothroyd et al. | The hearing aid input: A phonemic approach to assessing the spectral distribution of speech | |
| CN109151692A (en) | Hearing aid based on deep learning network tests method of completing the square certainly | |
| CN105307093B (en) | A kind of adaptive hearing compensation method | |
| CN108198571A (en) | A kind of bandwidth expanding method judged based on adaptive bandwidth and system | |
| CN106851511A (en) | A kind of quick pure tone test method based on smart mobile phone | |
| CN103892939B (en) | Improve language processing device for artificial cochlea and the method for Chinese tone recognition rate | |
| CN114339564B (en) | Neural network-based self-adaptation method for self-adaptive hearing aid of user | |
| CN113411733A (en) | Parameter self-adjusting method for fitting-free hearing aid | |
| CN116417146B (en) | CNN-based individual heart and lung function evaluation system and processing method thereof | |
| CN117116411A (en) | Method for constructing quantitative green land exposure physiological health effect threshold model | |
| CN108538301A (en) | A kind of intelligent digital musical instrument based on neural network Audiotechnica | |
| CN116453547A (en) | Hearing aid voice quality self-evaluation method based on hearing loss classification | |
| Shen et al. | Individualized estimation of the Speech Intelligibility Index for short sentences: Test-retest reliability | |
| CN109200469A (en) | A kind of Optimized Coding and system of the perception of enhancing artificial cochlea's tone |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221129 Address after: 518172 516, building 4, Qidi Xiexin Science Park, Qingchun Road, Longcheng street, Longgang District, Shenzhen, Guangdong Province Patentee after: SHENZHEN EARTECH Co.,Ltd. Address before: 300072 Tianjin City, Nankai District Wei Jin Road No. 92 Patentee before: Tianjin University |