Specific implementation mode
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Case is described in further detail.
Implementation environment
Referring to Fig.1, Fig. 1 is a kind of implementation environment configuration diagram of present example.
As shown in Figure 1, including:Host 101, user 102, sound generation source 103 to 109.
User 102 is the people for having hearing impairment or auditorily handicapped, has the ability on the head that freely swings, is helped wearing
After listening device, the instruction of dispenser can be followed, when hearing that around sound source makes a sound, can by way of swinging head,
Find sound source (i.e. facial face sound generation source direction).
Sound generation source 103 to 109 is sound generation source of the position distribution around user 102, and camera is mounted in test environment,
Optionally, may be mounted in sound generation source or user at the top of.In a kind of possible embodiment, sound generation source 103 to 109 with
Within the scope of 102 front 180 degree of family, it is evenly distributed on user 102 as the center of circle, on the circle of radius 3m, optionally, sound generation source 103,
It is equipped on 106 and 109 at the top of camera or user and camera is installed.
It should be noted that ensure that user is entirely to judge sound generation source position, user 101 and hair according to hearing
It needs to carry out visual isolation using barrier between sound source 103 to 109, as long as sound generation source can not be seen by reaching user 101, without
Influence hears that the sound effect that sound generation source is sent out is feasible measure, and optionally, all sound generation sources are placed in equipped with guard
It will on holder or be directly isolated etc. using opaque cloth between all sound generation sources and user 101, in addition, to meet one
As living scene needs, user 101 answers moderate at a distance from sound generation source in test environment, it is preferred that distance is 3m or so.
Host 101 can be the calculating such as desktop PC, notebook, palm PC and cloud server, smart mobile phone
Equipment etc..Host 101 can control any sound generation source of sound generation source 103 to 109 and play sound, and user 102 is enable to hear.In host
Corresponding debugging software is installed on 101, the relevant parameter of hearing aid can be adjusted, wherein the related ginseng of hearing aid
Number is primarily referred to as gain parameter in the present invention.
The hearing aid that host 101 is worn with user 102 is attached by wired or wireless way.A kind of possible
In embodiment, host 101 and user 102 pass through bluetooth connection.In alternatively possible embodiment, host 101 and use
Family 102 is connected by hearing aid line program.
Host 101 is connect with sound generation source 103 to 109 by wired or wireless way.
Skilled person will appreciate that above-mentioned implementation environment is only schematical explanation, do not constitute to the implementation environment
Specific restriction.
As Fig. 2 be the embodiment of the present invention in binaural hearing aid balance adjustment method one embodiment schematic diagram, including:
In step 201, testing with condition for balance adjustment is provided, tests and includes with condition:Fixed multiple sound generation sources, and really
The place angle of the fixed sound generation source.
The method of binaural hearing aid balance adjustment proposed by the invention be include certain implementation condition, i.e., specific
When operation, dispenser needs to meet in the place for carrying out hearing aid balance adjustment equipped with fixed sound generation source, here " Gu
It is fixed " it is emphasised that the quantity of sound generation source and position should be known determination in adjustment process.
Optimal, sound generation source shares 7, is spaced 30 degree between each, is evenly distributed on the semi-circular front arc of user.This
Testing for sample is accurate enough with the resolution ratio of condition in all directions, enables to adjust fully to meet and arrives daily actual scene
The middle identification needed to audio direction, the quantity for being further added by sound generation source is adjusted and the promotion very little brought.
All sound generation source composition test sound generation source groups, it is for convenience of description, following to be illustrated with testing sound generation source group.
It should be noted that the angle of any two sound generation source position in test sound generation source group is all different, such as
The sound generation source 103 to 109 of above-mentioned implementation environment, which is evenly distributed over, to be fixed on user 102 as the center of circle, the 180 degree model of radius 3m
This point can be fully ensured that on the circle enclosed, but in fact should be as limit.
In step 202, the first step, control targe sound generation source plays sound, wherein target utterance source is selected at random.
Dispenser is during binaural hearing aid balance adjustment, it is desirable to provide the sound generation source that can be made a sound, for user
It is tested.In embodiments of the present invention, sound generation source may be selected to be speaker, and test sound generation source group is correspondingly described as testing speaker
Group.
To ensure the reliability and accuracy of binaural hearing aid balance adjustment, it should includes at least two to test in sound generation source group
Sound generation source.In a kind of possible realization, tests in speaker group and share 2 speakers, be attached with host, master can be received
The play instruction of machine plays specified audio files, and optionally, the format of audio files includes WMA, MP3, MIDI etc..Wherein,
Host 101 in host corresponding diagram 1.
In practical operation, dispenser entire balance adjustment test match during, it usually needs repeatedly repeatedly test adjust
Section could complete test target.Target utterance source refers to dispenser in this adjustment process, and sound is played by host computer control
Sound generation source, and to be dispenser select at random from the device for not completing balance adjustment in adjustment process before in the target utterance source
It selects.
In a kind of possible realization, test sound generation source group shares 3 speakers, and number consecutively is No. 1, No. 2 and No. 3.?
In adjusting for the first time, 3 speakers were not tested, and dispenser may be selected to survey in a speaker in No. 1 or No. 2 or No. 3
Examination is adjusted for the first time if No. 1 speaker has been selected to complete, and in second is adjusted, one in No. 2 or No. 3 may be selected in dispenser
It is tested in a speaker, if No. 3 speakers has been selected to complete second of adjusting, in third time is adjusted, dispenser's selection is surplus
Under No. 2 speakers tested.
In step 203, second step determines the head angle of user.
In step 202, user, can will be facial under the instruction of dispenser after hearing that target utterance source plays sound
The position of object-oriented sound generation source.The quantity of camera in this step 203 should be not less than 3, just can guarantee by taking the photograph in this way
As user's head image information that head takes can be used for determining user's head angle enough.
In step 204, third step, according to the place angle calculation of the head angle of the user and the target utterance source
Angular deviation determines that target gain is poor.
After determining head angle of user after hearing sound by step 203, according to user's head angle and currently
Play the deviation between the angle of the device position of sound, it may be determined that the target for instructing hearing aid to adjust can be helped to increase
Benefit is poor.
In step 205, the 4th step is poor according to the target gain, and the hearing aid worn to the user carries out corresponding
Parameter regulation.
By target gain value determined by step 204, the parameter regulation for instructing to adjust hearing aid can be helped so that help
Listen device after adjustment, user's head angle can be closer to the real angle in target utterance source.By repeatedly adopting successively
With the first step, the method for the second step, the third step and the 4th step, the balance up to completing binaural hearing aid
It adjusts
In the embodiment of the present invention, a kind of method of binaural hearing aid balance adjustment is provided, first by host computer control target
Sound generation source plays sound, wherein target utterance source is that host is randomly choosed from test sound generation source group, and then, user is hearing
After target utterance source plays sound, under the instruction of dispenser, can by the position of the object-oriented sound generation source of face, then, according to
The angular deviation of user's head angle and target utterance source determines that target gain is poor, finally, poor according to target gain, to user
The hearing aid worn carries out corresponding parameter regulation, completes binaural hearing aid balance adjustment.By the above-mentioned means, dispenser's energy
Enough efficiently easily user of the guidance with hearing impairment or auditorily handicapped completes the test of binaural hearing aid balance adjustment, realizes double
The purpose of ear hearing aid balance adjustment, it is ensured that user, can be more in routine work life after wearing the hearing aid regulated
Accurately distinguish the source of sound.
For being to pass through random manner control targe sounding in step 202 in the first step in the above embodiment of the present invention
Source plays sound, is constantly tested, and the balance adjustment to be finally completed binaural hearing aid is feasible scheme, but inevitable
Longer length of testing speech is needed just to can ensure that all sound generation sources all in equilibrium state.
Therefore, to further increase balance adjustment efficiency, in another embodiment of the present invention, in above-mentioned steps,
Host computer control target speaker plays sound, wherein target speaker be main machine from test speaker group randomly selected process into one
Step optimizes, and referring to Fig. 3 and Fig. 4, detailed description are as follows:
Fig. 3 is the operation chart of main-engine control desk screen during binaural hearing aid balance adjustment.Fig. 4 is host
Control targe speaker plays the method flow diagram of sound.
In the present embodiment, sound generation source is selected as speaker, and in figure 3, main-engine control desk screen 301 is dispenser to helping
Device is listened to be balanced the operation interface in adjustment process, the host 101 in host corresponding diagram 1 here, speaker 302 to 308 is total to
With composition test speaker group.
Process described in Fig. 4 includes:
In step 401, it is grouped according to the equilibrium state of each sound generation source in the test sound generation source group, the test
Sound generation source group is divided into balance group and uneven group;
It should be noted that all sound generation sources in the balance group are in equilibrium state, and in the uneven group
All sound generation sources are in non-equilibrium state.Equilibrium state mentioned here is not the attribute of sound generation source itself, but for weighing
Whether amount sound generation source completes test during binaural hearing aid balance adjustment, if sound generation source does not complete balance test, says
The bright sound generation source is in non-equilibrium state, conversely, illustrating that the sound generation source is in equilibrium state.
In step 402, target utterance source is randomly choosed from uneven group in test sound generation source group;
It should be noted that random selection described in above-mentioned steps 402 simply to illustrate that target utterance source choosing
The difference for selecting sequence is unrestricted in the present invention, and does not lie in the uncertainty for emphasizing selecting sequence.
As shown in figure 3, the best embodiment of the use in the present embodiment is, tests in speaker group and shares 7 speakers,
1 to No. 7 shown on the main-engine control desk screen 301 entity speaker just corresponded in implementation environment, when speaker was matched testing
When equilibrium state in journey changes, the equilibrium state of entity speaker can be distinguished by background shape, i.e., whether is completed
Balance adjustment.Specifically, the round background of speaker 302 and speaker 303, number is 1 and 2 in characterization test speaker group entity sound
Balance adjustment is completed in case, is in equilibrium state, and the entity speaker that remaining number in testing speaker group is 3 to 7 is not yet
Balance adjustment is completed, background shape is rectangular, is in non-equilibrium state.So, in testing speaker group, 2 groups of speaker 1 and speaker
At balance group, speaker 3 to 7 constitutes uneven group.During this binaural hearing aid balance adjustment shown in Fig. 3, test
The specified audio files of the broadcasting of the speaker 3 in uneven group is controlled by clicking speaker 3 on main-engine control desk screen with teacher, this
When, speaker 3 is target speaker.It is understood that the balance group and uneven group in test speaker group can match process with testing
Progress and dynamic change.
Dispenser, which is further increased, compared to the method for a upper embodiment, in the present embodiment passes through host computer control target sound
Case plays the process of sound, is controlled test speaker group by the way that main-engine control desk is unified, can be square by visualization interface
Just testing progress is managed, such as dispenser can clearly know according to the background shape of correspondent entity speaker on interface
Road is in the speaker quantity of equilibrium state, i.e. balance group speaker quantity.
On the basis of above-mentioned Fig. 2 corresponding embodiments, the embodiment of the present invention determines the head angle of user, including:
Based on user's head portrait that camera takes, the head angle of the user is determined, wherein camera quantity is one
It is a or positioned at the multiple of different angle;
In method disclosed by the invention, camera quantity should be at least 1, for shooting the head of user during the test
Picture, the position of camera are required to accurately take the head portrait of user.In a kind of possible realization, the number of camera
Amount is 1, be may be mounted on the crown or any test sound generation source of user.In alternatively possible realization, camera
Quantity is 3, can be separately mounted to speaker 103 in Fig. 2, on speaker 106 and speaker 109.It should be noted that camera
When more, user's head portrait can be taken from more perspective, can be got the more head angle information of user, be contributed to
More accurately judge the head angle of user.
Based on user's head portrait that camera takes, before the head angle for determining the user, can also include:
The test position in user is obtained, head is in different angle, corresponding head photo that camera takes;
By carrying out characteristic extraction to head photo, and combine the corresponding head angle of head photo, using based on
The modeling method of data-driven establishes head angle decision model, wherein the input parameter of head angle decision model is head
The characteristic of photo, output parameter are head angle.
It should be noted that being balanced in the method for adjusting testing with during, a very important step is to want
It can determine user's head angle.Dispenser will be daily many different users, carry out a large amount of hearing aid fitting work,
Binaural hearing aid balance adjustment is the part tested during matching.Therefore, the invention discloses a kind of head angles to judge mould
Type acquisition methods, dispenser promptly can test the determining user's head angle with during using the model, substantially increase
Test the efficiency matched.
The head angle decision model acquisition methods that the present invention announces are based on the test position in user, and head is in difference
Great amount of samples head photo when angle, and by the characteristic of head portrait feature extraction algorithm acquisition head photo, and combine
The head angle of these head photos forms a large amount of sample data, using the modeling method based on data-driven, establishes user
Relationship between head portrait and user perspective.Optionally, the modeling method based on data-driven can be based on convolutional neural networks
(CNN) deep learning algorithm, the embodiment of the present invention will be used as explanation in this way, but the present invention is not limited to this sides
Method all has feasibility based on the modeling method of data-driven in scheme proposed by the present invention.
The present embodiment illustrates on the basis of Fig. 3 corresponding embodiments, and in conjunction with Fig. 5, and Fig. 5 is to obtain user's head angle
The schematic diagram of degree.
Fig. 5 includes camera 501 to 503, and user's head portrait 504 to 506, user's head portrait 504 is shot by camera 501
It obtains, user's head portrait 505 is obtained by the shooting of camera 502, and user's head portrait 506 is obtained by the shooting of camera 503.It is real
Apply and share 7 speakers in environment, camera is mounted on No. 1, No. 3 and No. 7 speaker, be corresponding in turn to camera 501 in Fig. 5 to
503。
The head portrait sample data of a large amount of different groups of acquisition first, is clapped specifically, one group of sample data should include camera
User's head portrait for taking the photograph and at this time user's head angle, that is, user's head towards angle.Then, using feature extraction
Algorithm carries out feature extraction to the head image data of user, retain face mask, eyes, nose and face location information, filtering
Additional detail is obtained the characteristic of head portrait, and the deep learning algorithm based on convolutional neural networks (CNN), is adopted using these
The a large amount of sample data collected is trained, and sets certain training error threshold value, obtains with user's head portrait being that input is joined
Number, using the head angle of user as the head angle decision model of output parameter.
In a kind of possible realization, using model or true man when head is turned to different angle on user location, note
Lower 507 head of user's head portrait 504,505 and 506 and user taken using No. 501, No. 502 and No. 503 cameras of record
Angle, such great amount of samples data can be as the training data of acquisition head angle decision model, and sets head angle
Error threshold, optional threshold value are set as 5 degree.When the head angle decision model obtained by training is applied in actual ears
When the method for hearing aid balance adjustment, the judgement of user's head angle does not exceed 5 degree, has higher precision, disclosure satisfy that
The actual demand of user.
In the present embodiment, it is contemplated that dispenser will be daily many different users, carry out a large amount of hearing aid fitting
Work, binaural hearing aid balance adjustment are the part tested during matching, and the present invention provides a kind of head angles to judge mould
Type acquisition methods, the model are the bases of the user's head portrait quick obtaining user's head angle taken by camera.
In a kind of possible realization, for determine user head angle camera only there are one, which can
With at the top of the user or position that other can be clearly captured user's head portrait, in the use taken based on a camera
Account picture before the head angle for determining the user, still may be used the head angle described in above-described embodiment and sentence
Cover half type acquisition methods, the characteristic that structure input is extracted for user's head portrait that the camera takes through image processing techniques
According to, export the head angle decision model for the head angle of the user corresponding to user's head portrait, for hearing aid balance
Head angle in adjusting for determining user.
The present embodiment is by camera to acquire user's head portrait and by (or other known skills in a manner of above-mentioned image recognition
The image recognition technology of art) obtain user head angle.But in fact, other modes can also be used, which to acquire, obtains user
Head angle, such as in the case where not considering implementation cost and efficiency, the helmet-type with angular transducer can also be used
The more acurrate head angle for obtaining user of head angle collecting device.
In yet another embodiment of the present invention, in Fig. 2 above-mentioned steps, according to user's head angle and target utterance source
Angular deviation, determine that the process of target gain difference is introduced.
By 203 in the step in Fig. 2, the user's head portrait taken by camera determines user's head angle, and
The angle in target utterance source is pre-determined, is given value.The present invention announces a kind of according to user's head angle and target
The angular deviation of sound generation source, the method for determining target gain difference.
In Fig. 6 of the present embodiment, a kind of optionally coordinate diagram about the definition of angle is shown, suitable for using account
Portion's angle and the angle in target utterance source.As shown in fig. 6, the position where user 601 is o points, using o points as pole, with ox
For polar axis, the positive direction clockwise for angle is taken, polar coordinate system, following user's head angles and target utterance source are established
Angle all in figure 6 shown in coordinate system measured.
It should be noted that the angular deviation between the head angle and the angle in target utterance source of user refers to user
Head angle is to the angular deviation in target utterance source, and when deviation is clockwise direction, angular deviation is just, when deviation is the inverse time
When needle direction, angular deviation is negative.In a kind of possible realization, the head angle of user is(120 degree), target utterance
The angle in source is(90 degree), then angular deviation is(- 30 degree).
What target gain difference reflected is the regulated quantity of hearing aid gain, and unit db passes through the gain for adjusting hearing aid, energy
The auditorily handicapped for enough making up different type or degree, when hearing aid to auditorily handicapped make up effect it is better when, testing match process
In, the angular deviation in user's head angle and target utterance source should be smaller, therefore user's head angle and target utterance
It is to have certain relationship that the angu-lar deviation in source, which is between target gain difference,.
The present invention discloses between a kind of user's head angle and the angu-lar deviation x and target gain difference g in target utterance source
Relationship meet following relationship:
Wherein, the domain of x is [- π, π], wherein positive value indicates misalignment angle toward clockwise, negative value is counterclockwise
Direction.The relationship shows the absolute value of the angu-lar deviation in user's head angle and target utterance source when less than 45 degree, with
The relationship being proportionate between target gain difference, when more than or equal to 45 degree, target gain difference be 10, when less than or equal to-
At 45 degree, target gain difference is -10, ranging from -10~the 10 of target gain difference, unit db.
In a kind of possible realization, as shown in fig. 6, sound generation source 602 is speaker, place angle isIt may at one group
The determination of the target gain difference of appearance is as shown in table 1.
It in the present embodiment, according to the angular deviation of user's head angle and target utterance source, determines that target gain is poor, is right
The gain of hearing aid is adjusted, and the position for enabling users to more accurately resolution target sound generation source lays the foundation.
Based on the embodiment corresponding to Fig. 2, hearing aid that the present embodiment is worn to poor according to target gain, to user into
The corresponding parameter regulation of row, the process for completing binaural hearing aid balance adjustment are introduced.
It is different can so that hearing aid brings the user with auditorily handicapped by the gain of adjusting hearing aid by dispenser
Sound capabilities are distinguished, including the source for distinguishing sound.In the method disclosed in the present, when target gain difference it is absolute
Dispenser just needs that the gain of hearing aid is adjusted when value is more than preset threshold range, illustrates the parameter of hearing aids today
User cannot be met by, which being arranged, distinguishes that sound source is tested with target, that is to say, that the hearing aid balance for not up to adjusting user needs
Hearing aid gain is further adjusted, optionally, the predeterminable threshold value of target gain absolute value of the difference is 1.
It is understood that in ideal conditions, it is dispenser that the device quantity tested in sound generation source group is corresponding needs
Will be to the maximum possible number that the gain of hearing aid is adjusted, in practical operation, the speaker quantity tested in speaker group is
7, illustrate that dispenser needs to test user using the speaker in different angle, i.e., test 7 times is needed altogether, in one kind
In possible realization, in the 1st time to the 6th time test, target gain absolute value of the difference is above preset threshold value 1, and the 7th
In secondary test, target gain absolute value of the difference in threshold value 1, then, hearing aid be required in being tested at first 6 times carry out gain ginseng
Number is adjusted, and the 7th time does not need then, in alternatively possible realization, in the 1st time to the 6th time test, target gain difference
Absolute value is above preset threshold value 1, and in being tested at the 7th time, target gain absolute value of the difference is still above threshold value 1, then helping
It is required for carrying out gain parameter adjusting in listening device to test at 7 times, to complete binaural hearing aid balance adjustment.In other cases,
After the step of completing 1 speaker corresponding balance adjustment in test speaker group, other speakers pair in testing speaker group
After the step of balance adjustment answered, originally completed balance adjustment step correspondence speaker may occur it is unbalanced again can
Can, then it also needs to be balanced regulating step again.In short, the step of by repeatedly testing the balance adjustment of speaker in speaker group,
Until all speakers in entirely test speaker group are all in equilibrium state.
The present invention announces a kind of method for the gain that target gain difference can instruct dispenser how to adjust hearing aid, that is, works as
The method that target gain absolute value of the difference determines the regulated quantity of hearing aid gain in threshold range.
In following for convenience of description, target gain difference is indicated with g, and the preset threshold value of target gain difference is indicated with t, when
g>When t, g/2 is increased to the left ear gain of hearing aid, auris dextra hearing aid reduces g/2, works as g<When-t, the left ear gain to hearing aid
Reduce g/2, auris dextra hearing aid increases g/2, and when g is in range-t~t, hearing aid need not be adjusted in explanation.Wherein, g
Ranging from -10~10, the mode specifically obtained has been described above in upper one embodiment, and which is not described herein again, t be more than 0
Positive number.
In a kind of possible realization, t is preset as 1db, the hearing aid gain that one group is likely to occur regulated quantity really
Recognize as shown in table 2.
Target gain difference g (db) |
Left ear hearing aid gain regulated quantity (db) |
Auris dextra hearing aid gain regulated quantity (db) |
10 |
5 |
-5 |
1 |
Without adjusting |
Without adjusting |
-1 |
Without adjusting |
Without adjusting |
-10 |
-5 |
5 |
Table 2
It should be noted that as shown in table 2, as g=10, in conjunction with a upper embodiment, it is known that user's head angle at this time
Angular deviation with target utterance source is more than or equal to(45 degree), as g=1, angular deviation at this time is about 0.08rad (4.5
Degree), hearing aid need not be adjusted, illustrate that the parameter of hearing aid at this time can meet user's discrimination objective sounding source of sound
Direction precision.
In the present embodiment, one kind is provided in binaural hearing aid balance adjustment, determines the regulated quantity of hearing aid gain
Method, while by it is repetitious adjust so that hearing aid gain parameter can increasingly meet allow dispenser complete ears
Testing with target for hearing aid balance adjustment, a kind of effective and feasible method is proposed for binaural hearing aid balance adjustment.
You need to add is that above-described embodiment illustrates user's head angle and target utterance source with a concrete condition
The adjustment process of angu-lar deviation x and target gain difference g are only used for showing a kind of specific implementation process of the embodiment, wherein
The parameter of relationship between user's head angle and the angu-lar deviation x and target gain difference g in target utterance source and preset
Threshold value is can be adjusted and change with condition and actual needs according to for testing for balance adjustment, should not be limited with this
Protection scope of the present invention.
On the basis of Fig. 2 institutes embodiment, the present embodiment by conjunction with Fig. 7 come to disclosed in this invention binaural hearing aid
The method of balance adjustment carries out more complete specific description.Fig. 7 is the method flow diagram of binaural hearing aid balance adjustment.
Sound generation source in the present embodiment is selected as speaker and illustrates, with the embodiment corresponding to Fig. 2.
In step 701, all speakers are labeled as non-equilibrium state;
In step 702, the speaker of a non-equilibrium state is randomly choosed, plays sound;
In step 703, according to user's head portrait of the camera shooting on 1,4 and No. 7 speaker, mould is judged by head angle
Type determines user's head angle;
In step 704, user's head angle is compared with sounding speaker angle, calculates angular deviation;
In step 705, according to angular deviation, determine that target gain is poor;
In step 706, if target gain difference is less than preset threshold value, gos to step 708, otherwise enter step 707;
Poor according to target gain in step 707, the hearing aid worn to user carries out corresponding parameter regulation;
In step 708, target speaker is set to equilibrium state;
In step 709, if there is also the speaker of non-equilibrium state, 702 are gone to step, it is on the contrary then illustrate that ears help
Device balance adjustment is listened to test completion.
It should be noted that it is a kind of ideally, according to the uneven speaker in different angle to hearing aid into
During row test and adjusting, do not conflict to the regulating effect of hearing aid gain.
In a kind of possible realization, dispenser selects No. 1 speaker to test first, according to above-mentioned steps to hearing aid
Gain be adjusted, detailed process repeats no more so that when hearing that No. 1 speaker plays sound, No. 1 speaker position
Angular deviation between angle and user's head angle is in set threshold range, then No. 1 speaker can be marked as equilibrium-like
State, then, dispenser continue to select in uneven voice box group, optionally, No. 3 speakers are selected to play sound, still root
The gain of hearing aid is adjusted according to above-mentioned steps so that when hearing that No. 3 speakers play sound, No. 3 speaker positions
Angle and user's head angle between angular deviation in set threshold range, at the same time, No. 1 speaker should be still in flat
Weighing apparatus state.
An embodiment of the present invention provides a kind of methods of binaural hearing aid balance adjustment, first, control targe sound generation source hair
Go out sound, user swings head, user's head angle is judged according to user's head portrait that camera takes, according to target utterance source
Hearing aid gain parameter is adjusted in angular deviation between place angle and user's head angle, until target utterance source
It is balance by target utterance source marking until angular deviation between place angle and user's head angle is less than preset threshold value
State, then tests next speaker in non-equilibrium state, and equilibrium state is in until all speakers, this
Sample make user can reach after wear hearing aid ears balance, compensate for test with do not fully considered in link user's ears balance
The defect of adjusting provides to the user and more improves effective test with service.
Following is apparatus of the present invention embodiment, can be used for executing the method for the present invention embodiment.For apparatus of the present invention reality
Undisclosed details in example is applied, the method for the present invention embodiment is please referred to.
It please refers to shown in Fig. 8, it illustrates the devices of binaural hearing aid balance adjustment provided by one embodiment of the present invention
Block diagram.The device, which has, executes the exemplary function of the above method, and function can also be executed by hardware realization by hardware
Corresponding software realization.The device may include:
Fixed multiple sound generation sources;
Playing module 801 is controlled, plays sound for control targe sound generation source, wherein target utterance source is selected at random;
First determining module 802, the head angle for determining user;
Second determining module 803, according to the place angle calculation of the head angle of the user and the target utterance source
Angular deviation determines that target gain is poor;
Adjustment module 804, for poor according to the target gain, the hearing aid worn to the user carries out accordingly
Parameter regulation.
Optionally, playing module 801 is controlled, is used for:
It is described to test with the sound generation source composition test sound generation source group in condition, according to each in the test sound generation source group
The equilibrium state of sound generation source is grouped, and the test sound generation source group is divided into balance group and uneven group, wherein the balance group
In all sound generation sources be in equilibrium state, and all sound generation sources in the uneven group are in non-equilibrium state;
The target utterance source is randomly choosed from uneven group in test sound generation source group.
Optionally, the first determining module 802, is used for:
The test position in the user is obtained, head is in different angle, corresponding head that the camera takes
Portion's photo;
By carrying out characteristic extraction to the head photo, and in conjunction with the corresponding head angle of head photo,
Head angle decision model is established using the modeling method based on data-driven, wherein the input of head angle decision model is joined
Number is the characteristic of the head photo, and output parameter is the head angle.
Optionally, the second confirmation module 803, is used for:
When the absolute value of the angular deviation is no more than preset threshold range, the target gain difference and the angle
The relationship that deviation is proportionate;
When the absolute value of the angular deviation is more than preset threshold range, the value of the target gain difference is constant.
Optionally, adjustment module 804 are used for:
When target gain absolute value of the difference is more than preset threshold range, the gain to the hearing aid is needed to adjust
Section.
In the present embodiment, sound is played by host computer control target utterance source first, wherein target utterance source is host from survey
It is randomly choosed in examination sound generation source group, then, user is after hearing that target utterance source plays sound, under the instruction of dispenser, meeting
The position of the object-oriented sound generation source of face then according to the angular deviation of user's head angle and target utterance source, is determined into mesh
Gain inequality is marked, finally, poor according to target gain, the hearing aid worn to user carries out corresponding parameter regulation, completes ears
Hearing aid balance adjustment.By the above-mentioned means, dispenser efficiently can easily instruct to have hearing impairment or auditorily handicapped
User completes the test of binaural hearing aid balance adjustment, realizes the purpose of binaural hearing aid balance adjustment, it is ensured that user is wearing tune
After the hearing aid saved, the source of sound can be more accurately distinguished in routine work life.
Further, the process that dispenser plays sound by host computer control target speaker is increased, host computer control is passed through
Platform is unified to control test speaker group, can be easily managed to testing progress by visualization interface, such as test
The speaker quantity in equilibrium state can be known clearly according to the background shape of correspondent entity speaker on interface with teacher, i.e., it is flat
Weighing apparatus group speaker quantity.
In view of dispenser will be daily many different users, a large amount of hearing aid fitting work is carried out, ears help
It is the part tested during matching to listen device balance adjustment.The present invention provides a kind of head angle decision model acquisition methods,
The model is the basis that user's head angle is obtained by user's head portrait that camera takes.
According to the angular deviation of user's head angle and target utterance source, determine that target gain is poor, for the increasing to hearing aid
Benefit is adjusted, and the position for enabling users to more accurately resolution target sound generation source lays the foundation.
One kind is provided in binaural hearing aid balance adjustment, is adjusted by repetitious so that the gain of hearing aid is joined
Number, which can increasingly meet, to be allowed dispenser to complete binaural hearing aid balance adjustment to test with target, be embodied as ears effective and feasiblely and help
Listen device balance adjustment.
Fig. 9 shows the structural schematic diagram of the host involved by an example embodiments of the invention.The host includes:Place
Manage device 901, memory 902 and bus 903.
Processor 901 includes one or more processing cores, and processor 902 passes through bus 903 and 901 phase of processor
Even, memory 903 realizes above-mentioned implementation for storing program instruction, processor 901 when executing the program instruction in memory 902
The method for the binaural hearing aid balance adjustment that example provides.
Optionally, memory 902 can be by any kind of volatibility or non-volatile memory device or their group
It closes and realizes, such as static access memory (SRAM), electrically erasable programmable read-only memory (EEPROM) at any time are erasable to compile
Journey read-only memory (EPROM), programmable read only memory (PROM), read-only memory (ROM), magnetic memory, flash
Device, disk or CD.
The present invention also provides a kind of computer readable storage medium, be stored in the storage medium at least one instruction,
At least one section of program, code set or instruction set, at least one instruction, at least one section of program, code set or instruction set are by institute
The method for stating the binaural hearing aid balance adjustment that processor is loaded and executed to realize above method embodiment offer.
Optionally, the present invention also provides a kind of computer program products including instruction, when it runs on computers
When so that the method that computer executes the binaural hearing aid balance adjustment described in above-mentioned various aspects.
One of ordinary skill in the art will appreciate that realizing that all or part of step of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store computer-readable with one kind
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, be not used to limit invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.