CN108335704A - Vagitus detection circuit and relevant detection method - Google Patents
Vagitus detection circuit and relevant detection method Download PDFInfo
- Publication number
- CN108335704A CN108335704A CN201710044541.8A CN201710044541A CN108335704A CN 108335704 A CN108335704 A CN 108335704A CN 201710044541 A CN201710044541 A CN 201710044541A CN 108335704 A CN108335704 A CN 108335704A
- Authority
- CN
- China
- Prior art keywords
- sound
- signal
- sound clip
- vagitus
- circuit
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 59
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 238000005070 sampling Methods 0.000 claims description 14
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 6
- 230000008030 elimination Effects 0.000 claims 1
- 238000003379 elimination reaction Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 6
- 230000035945 sensitivity Effects 0.000 description 4
- 206010011469 Crying Diseases 0.000 description 3
- 238000000034 method Methods 0.000 description 3
- 239000012634 fragment Substances 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 241000638935 Senecio crassissimus Species 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/21—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being power information
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Hospice & Palliative Care (AREA)
- Psychiatry (AREA)
- General Health & Medical Sciences (AREA)
- Child & Adolescent Psychology (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention relates to a kind of vagitus detection circuit and relevant detection method, include a signal acquisition circuit, a feature extraction circuit and a decision circuitry, wherein the signal acquisition circuit is to when the intensity of a voice signal is more than a critical value, the voice signal is captured to generate a sound clip signal, wherein the time span of the corresponding sound clip of the sound clip signal is within a particular range；This feature acquisition circuitry is coupled to the signal acquisition circuit, and to capture multiple characteristic values of the sound clip signal；And the decision circuitry is coupled to this feature acquisition circuitry, and to according to those characteristic values come whether judge the sound clip corresponding to the sound clip signal be vagitus.
Description
Technical field
Present invention is directed to sound detection, espespecially a kind of vagitus detection circuit and relevant detection method.
Background technology
Current baby's monitor is typically according to the intensity size of received sound to determine whether there is baby cried
Sound occurs, for example, it is fixed critical that baby's monitor may determine that whether the intensity of received voice signal is more than one
Value, judges that the voice signal is vagitus, and send out alarm signal to parent if being greater than the critical value.However, above-mentioned
Judge whether voice signal is that the method for vagitus is possible to be influenced by ambient sound using critical value, thus causes
Erroneous judgement.
Invention content
Therefore, one of the objects of the present invention is to provide a kind of vagitus detection circuit and relevant detection method,
Multiple sound clips can be generated to be segmented to the voice signal received with reference to the crying characteristic of baby, and to each
A sound clip carries out characteristic value acquisition and comparison, accurately to judge whether received voice signal is baby cried
Sound, to solve the problems in prior art.
In one embodiment of the invention, a kind of vagitus detection circuit is disclosed, it includes have a signal acquisition
Circuit, a feature extraction circuit and a decision circuitry, wherein the signal acquisition circuit is to big in the intensity of a voice signal
When a critical value, which is captured to generate a sound clip signal, the wherein sound clip signal corresponds to
A sound clip time span within a particular range；This feature acquisition circuitry is coupled to the signal acquisition circuit,
And to capture multiple characteristic values of the sound clip signal；And the decision circuitry couples this feature acquisition circuitry, and use
With according to those characteristic values come whether judge the sound clip corresponding to the sound clip signal be vagitus.
In another embodiment of the present invention, a kind of vagitus detection method is disclosed, it includes have：In a sound
When the intensity of signal is more than a critical value, which is captured to generate a sound clip signal, the wherein sound
The time span of the corresponding sound clip of fragment signal is within a particular range；Capture the more of the sound clip signal
A characteristic value；And according to those characteristic values come whether judge the sound clip corresponding to the sound clip signal be baby cried
Sound.
Description of the drawings
Fig. 1 is the block diagram of the vagitus detection circuit of one embodiment of the invention.
Fig. 2 is the block diagram of the pretreatment circuit of one embodiment of the invention.
Fig. 3 is that signal acquisition circuit carries out segmentation acquisition to generate the schematic diagram of sound clip signal to voice signal.
Fig. 4 is the block diagram of the feature extraction circuit of one embodiment of the invention.
Fig. 5 is characterized multiple sound frames in acquisition circuitry and multiple characteristic parameters corresponding to it and multiple characteristic values
Example.
Fig. 6 is the flow chart according to the vagitus detection method of one embodiment of the invention.
Specific implementation mode
Referring to FIG. 1, it is the block diagram according to the vagitus detection circuit 100 of one embodiment of the invention.Such as Fig. 1 institutes
Show, vagitus detection circuit 100 contains a pretreatment circuit 110, a signal acquisition circuit 120, a feature extraction circuit
130, feature scaling circuit 140, a sound clip signal judging circuit 150 and a voice signal decision circuitry 160.At this
In embodiment, vagitus detection circuit 100 can be set in any electronic device for detecting vagitus, and should
Electronic device is used for being placed in the environment residing for baby, and a police is sent through wireless transmission when detecting vagitus
Show signal to another electronic device, to notify parent or caretaker.
In vagitus detection circuit 100, pretreatment circuit 110 is used for locating the voice signal received in advance
Reason.Specifically, referring to FIG. 2, it depicts the block diagram of the pretreatment circuit 110 of one embodiment of the invention, it comprises
One sampling frequency conversion circuit 210, a noise canceller circuit 220 and a gain circuitry 230.Since different vagitus is examined
The voice signal that slowdown monitoring circuit 100 is received may be for different frequencies or comprising a variety of different frequencies, in order to adapt to difference
Vagitus detection circuit 100, therefore sampling frequency conversion circuit 210 can be sampled frequency to the voice signal received
Conversion, such as using a fixed sampling frequency (for example, 8kHz) voice signal is sampled to generate sampling frequency
Voice signal after rate conversion.In another embodiment, specific vagitus detection circuit 100 can also be directly selected, this
When pretreatment circuit 110 can not need sampling frequency conversion circuit 210.Noise canceller circuit 220 is used for the sampling frequency
Voice signal carries out noise Processing for removing to generate voice signal after noise is eliminated after conversion.Gain circuitry 230 is used for making an uproar to this
Voice signal carries out Gain tuning processing after sound is eliminated, to generate voice signal after a pretreatment.In realization, noise canceller circuit
220 can be interchanged with the sequence of gain circuitry 230.If in addition, can tolerate poor treatment effect, gain circuitry 230 can be by
It removes.
It is shown in FIG. 1 pretreatment circuit 110 be one can be removed (optional) element, that is, the present invention another
In embodiment, pretreatment circuit 110 can be removed from vagitus detection circuit 100, and signal acquisition circuit 120 directly connects
Receive the voice signal.
With continued reference to FIG. 1, signal acquisition circuit 120 is used for capturing wherein one section letter of voice signal after the pretreatment
Number.Specifically, whether the intensity of voice signal is more than a critical value after the detection of acquisition circuitry 120 pretreatment, and when detection
When the intensity of voice signal is more than the critical value after to the pretreatment, voice signal after the pretreatment is captured and is somebody's turn to do with obtaining
A wherein sound clip signal for voice signal after pretreatment, the sound clip signal correspond to a sound clip, and the sound piece
The time span of section is within a particular range.In the present embodiment, the characteristic based on vagitus, the particular range between
Between 0.5~3 second.Specifically, referring to FIG. 3, after signal acquisition circuit 120 detects the pretreatment voice signal it is strong
When degree is more than the critical value, signal acquisition circuit 120 starts to capture voice signal after the pretreatment, until the pretreatment
The intensity of voice signal has arrived at the upper limit of the particular range (for example, this implementation less than the critical value or acquisition time afterwards
3 seconds in example), to generate a sound clip signal.In another embodiment of the present invention, if sound after the pretreatment
(for example, being more than 3 seconds) is higher than the critical value to the intensity of signal for a long time, then signal acquisition circuit 120 can capture a sound
After tablet segment signal (sound clip that corresponding time span is 3 seconds), and then start again at immediately after capturing the pretreatment
Voice signal is to capture next sound clip signal.
Feature extraction circuit 130 is used for capturing multiple characteristic values of each sound clip signal.Specifically, it please refers to
The feature extraction circuit 130 of Fig. 4, one embodiment of the invention contain a pre-emphasis (Pre-emphasis) circuit 410, a sound frame
Change (framing) circuit 420, a window (window) function computation circuit 430, a Fourier transform circuit 440, Meier filtering
Device group 450, a discrete cosine transform circuit 460 and an analysis circuit 470.It is first in the operation of feature extraction circuit 130
First, pre-emphasis circuit 410 carries out high-pass filtering operation to the sound clip signal, to generate a pre-emphasis signal, wherein pre- strong
Adjust the operation of circuit 410 that can be used as example explanation using following formula：X ' [n]=x [n] -0.97x [n-1], wherein x [n]
For the input of pre-emphasis circuit 410, and the output that x ' [n] is pre-emphasis circuit 410.Due to voice signal from generating body (such as
Baby) in sounding to the receive process of radio equipment (such as vagitus detection circuit 100), high frequency section can be with frequency
The phenomenon that increasing and generating energy attenuation, therefore the decaying that can compensate for high frequency section, in other words, energy are operated through Gao Tong Filter waves
More highlight the formant of high frequency.Sound frame circuit 420 takes out multiple sound frames from the pre-emphasis signal, for example, sound frame
Circuit 420 takes out the sound frame that multiple time spans are 20~40 milliseconds (ms) from the pre-emphasis signal (a corresponding sound clip)
(each sound frame corresponds to multiple sample points), and in order to avoid the variation of adjacent two sounds frame is excessive, therefore enable adjacent sound frame each other
There is part overlapping.Then, it is multiple to generate respectively to be multiplied by a window function by window function counting circuit 430 respectively for multiple sound frame
Window function sound frame, the wherein operation of window function counting circuit 430 can be used as example explanation using following formula：Y [n]=
X ' [n] * w [n], wherein y [n] they are the output of window function counting circuit 430, and w [n] is window function, and in one embodiment, window letter
Number Specifically, the operation of sound frame circuit 420 is by signal
Processing has regular length for every sound frame, therefore is easily processed, but since the signal in sound frame remains original amplitude,
And the signal outside sound frame is then set to 0, therefore discontinuous problem is caused, and penetrate the behaviour of window function counting circuit 430
Aforementioned discontinuous problem is can effectively eliminate, such as through Hamming window (Hamming Window) function, can be retained
The value of part and oppressive signal both ends among signal, the overlapping of adjacent tone frame is coordinated using this characteristic, you can sound frame is made to exist again
Apparent non-continuous event is not had on boundary.Fourier transform circuit 440 to multiple window function sound frame carry out from
Dissipate Fourier transform to generate sound frame after multiple Fourier transforms, the wherein operation of Fourier transform circuit 440 can use with
Lower formula is used as example explanation：Then, Meier filter group 450 is to those Fu
Sound frame is filtered after vertical leaf conversion, and to generate sound frame after multiple filtering, the wherein operation of Meier filter group 450 can use
Following formula is used as example explanation： Specifically, Meier filter group 450
M triangle bandpass filter is contained, and those triangle bandpass filters are evenly distributed on mel-frequency and are listened to simulate human ear
Feel characteristic.The energy frequency spectrum of multiple window function sound frame after aforementioned Fourier transform is penetrated to M triangle band logical respectively to filter
After the filtering of wave device, the energy being distributed on each mel-frequency can be found out.Discrete cosine transform circuit 460 is to multiple filter
Sound frame carries out discrete cosine transform to generate the multiple characteristic parameters for corresponding to each sound frame (also that is, Mel-cepstral after wave
Coefficient).Finally, analysis circuit 470 generates the acquisition signal according to corresponding to multiple characteristic parameter of each sound frame
Multiple characteristic value.
Pre-emphasis circuit 410 and window function counting circuit 430 shown in Fig. 4 are the element that can be removed, that is, in this hair
In another bright embodiment, pre-emphasis circuit 410 and/or window function counting circuit 430 can be from feature extraction circuits 130
It removes.
Referring to FIG. 5, its be characterized multiple sound frames in acquisition circuitry 130 and multiple characteristic parameters corresponding to it with
The example of multiple characteristic values.With reference to figure 5, it is assumed that sound clip signal has been removed N number of sound frame, and each sound frame has 12
Characteristic parameter C1~C12, the at this time meeting of analysis circuit 470 count the characteristic parameter of the identical number of each sound frame respectively
It calculates, to obtain the median and interquartile range corresponding to each characteristic parameter C1~C12, that is, can obtain in 12
Digit and 12 interquartile ranges.In addition, above-mentioned 12 medians and 12 interquartile ranges, in addition 12 interquartile ranges
One root-mean-square valve can be used as 26 features to join along with the quantity (such as N) for the sound frame that the sound clip signal is removed
It counts using the output as feature extraction circuit 130.
With continued reference to FIG. 1, feature scales circuit 140 to the corresponding characteristic value of the same sound clip signal (before such as
26 characteristic values stated) operation is zoomed in and out to maintain the stabilization of numberical range, and generate characteristic value after scaling.Sound clip is believed
Number decision circuitry 150 is directed to the same sound piece according to one to machine algorithm (Support Vector Machines, SVM)
Characteristic value (such as characteristic value after 26 scalings above-mentioned) calculates to judge the sound clip after the corresponding scaling of segment signal
Whether the corresponding sound clip of signal is vagitus.And in one embodiment, which is basad with diameter
The Zhi Xiangji algorithms of function (Radial Basis Function, RBF) core.Specifically, engineer's meeting at factory end
Training data (training data) is first input to one into machine algorithm study module, it is one super flat to determine to be located at
Multiple supports on face (hyperplane) are vectorial (support vector), using as one to machine model, wherein the branch to
Machine model is that two set with maximum boundary (margin) are established in two dimensional surface；And in practical operation, sound piece
Segment signal decision circuitry 150 can judge (such as 26 contractings above-mentioned of characteristic value after the corresponding scaling of the same sound clip signal
Put rear characteristic value) which set belonged to, and judge whether the corresponding sound clip of sound clip signal is baby according to this
Crying.
In addition, feature scaling circuit 140 is the element to can be removed in itself, that is, in another embodiment of the present invention
In, feature scaling circuit 140 can be removed from vagitus detection circuit 100.
Voice signal decision circuitry 160 can be set according to a sensitivity, to be judged according at least one sound clip signal
The judging result of circuit determines whether the voice signal is vagitus.For example, when 100 quilt of vagitus detection circuit
When being set as having highly sensitive, as long as it is vagitus to be judged there are one sound clip signal, then voice signal judges electricity
Road 160 will determine that the voice signal is vagitus, vagitus detection circuit 100 send according to this alarm signal to parent or
It is caretaker；When vagitus detection circuit 100 is set with medium sensitivity, continuous 5 sound clip signals
In to have 2 sound clip signals to be judged be vagitus, voice signal decision circuitry 160 will determine that the voice signal is baby
Vagitus sound；And when vagitus detection circuit 100 is set with muting sensitivity, in continuous 5 sound clips at least
It is vagitus that 3 sound clip signals, which are judged, and voice signal decision circuitry 150 can just determine that the voice signal is baby
Vagitus sound.
Sound clip signal judging circuit 150 and voice signal decision circuitry 160 the two decision circuitries are set in Fig. 1
The reason of be that therefore in one embodiment, sound clip signal judging circuit 150 itself is i.e. available the problem of sensitivity is contemplated
Determine that the voice signal is vagitus, and voice signal decision circuitry 160 can be moved from vagitus detection circuit 100
It removes.In another embodiment, sound clip signal judging circuit 150 and voice signal decision circuitry 160 can be same
Carry out implementation in a circuit module.
Referring to FIG. 6, it is the flow chart according to the vagitus detection method of one embodiment of the invention.With reference to figure
The related narration of 1~5 embodiment, flow shown in fig. 6 are as described below.
Step 600：Flow starts.
Step 602：Whether the intensity for detecting a voice signal is more than a critical value, and works as and detect the strong of the voice signal
When degree is more than the critical value, which is captured to generate at least a sound clip signal, the wherein sound clip
The time span of the corresponding sound clip of signal is within a particular range.
Step 604：Calculate multiple characteristic values of the sound clip signal.
Step 606：Judge whether the sound clip signal is vagitus according to multiple characteristic value.
Step 608：It is to determine the voice signal according to whether the sound clip signal is the judging result of vagitus
No is vagitus.
The brief summary present invention is to refer to baby in the vagitus detection circuit and related methods of the present invention
Crying characteristic captures carrying out segmentation to the voice signal received and generates multiple sound clip signals, wherein each sound
The time span of fragment signal has a particular range, such as 0.5~3 second, is carried out again to each sound clip signal later
Characteristic value is captured and is compared, accurately to judge whether received voice signal is vagitus.It, can through the present invention
To reduce the influence of ambient sound really, the accuracy of vagitus detection and judgement is promoted.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification should all belong to the covering scope of the present invention.
Claims (24)
1. a kind of vagitus detection circuit, including：
One signal acquisition circuit, when the intensity of a voice signal is more than a critical value, to be captured to the voice signal
To generate a sound clip signal, wherein the time span of the corresponding sound clip of the sound clip signal is between a specific model
Within enclosing；
One feature extraction circuit is coupled to the signal acquisition circuit, to capture multiple characteristic values of the sound clip signal；With
And
One decision circuitry is coupled to this feature acquisition circuitry, to judge the sound clip signal pair according to those characteristic values
Whether the sound clip answered is vagitus.
2. vagitus detection circuit as described in claim 1, feature is in when the intensity of the voice signal is critical more than this
When value, which starts to capture the voice signal, until the intensity of the voice signal is less than the critical value
Or one the acquisition time have arrived at the upper limit of the particular range, to generate the sound clip signal.
3. vagitus detection circuit as claimed in claim 2, which is characterized in that if the signal acquisition circuit is because capture
When time has arrived at the upper limit of the particular range and generates the sound clip signal, the signal acquisition circuit is from the acquisition time
The time point for reaching the upper limit of the particular range starts to capture the voice signal to generate next sound clip signal.
4. vagitus detection circuit as described in claim 1, which is characterized in that the particular range be 0.5 second to 3 seconds it
Between.
5. vagitus detection circuit as described in claim 1, which is characterized in that also include：
One pretreatment circuit is picked to be pre-processed to the voice signal with generating voice signal to signal after a pretreatment
Sense circuit, and the pretreatment circuit includes：
One sampling frequency conversion circuit generates to use a fixed sampling frequency to be sampled to the voice signal
Voice signal after the conversion of one sampling frequency；
One noise canceller circuit is coupled to the sampling frequency conversion circuit, to the sampling frequency convert after voice signal into
Row noise Processing for removing is to generate voice signal after a noise is eliminated；And
One gain circuitry is coupled to the noise canceller circuit, to be carried out at Gain tuning to voice signal after noise elimination
Reason, to generate voice signal after the pretreatment.
6. vagitus detection circuit as described in claim 1, which is characterized in that this feature acquisition circuitry includes：
One sound frame circuit, to take out multiple sound frames from the sound clip signal；
One Fourier transform circuit, to carry out Fourier transform to those sound frames to generate sound frame after multiple Fourier transforms；
One filter group, be filtered to sound frame after those Fourier transforms to generate sound frame after multiple filtering；
One discrete cosine transform circuit corresponds to each to carry out discrete cosine transform to sound frame after those filtering to generate
Multiple characteristic parameters of sound frame；And
One analysis circuit, to according to corresponding to those characteristic parameters of each sound frame come generate the sound clip signal should
A little characteristic values.
7. vagitus detection circuit as claimed in claim 6, which is characterized in that this feature acquisition circuitry also includes：
One window function counting circuit, to handle those sound frames according to a window function to generate multiple window function sound frames, and should
Fourier transform circuit carries out Fourier transform to generate sound frame after those Fourier transforms to those window function sound frames.
8. vagitus detection circuit as claimed in claim 6, which is characterized in that this feature acquisition circuitry also includes：
One pre-emphasis circuit is used for carrying out high-pass filtering operation to the sound clip, to generate a pre-emphasis signal, and the sound frame
Change circuit and takes out those sound frames from the pre-emphasis signal.
9. vagitus detection circuit as described in claim 1, which is characterized in that this feature acquisition circuitry includes：
One sound frame circuit, to take out multiple sound frames from the sound clip signal, wherein those characteristic values correspond to those respectively
Sound frame, and the decision circuitry is sentenced according to the quantity of multiple medians of those characteristic values, multiple interquartile ranges and those sound frames
Whether the sound clip for the sound clip signal correspondence of breaking is vagitus.
10. vagitus detection circuit as described in claim 1, which is characterized in that the decision circuitry is drilled using one to machine
Whether algorithm is to judge the corresponding sound clip of the sound clip signal being vagitus according to those characteristic values.
11. vagitus detection circuit as claimed in claim 10, which is characterized in that the Zhi Xiangji algorithms are with radial direction
The Zhi Xiangji algorithms of basis function core.
12. vagitus detection circuit as described in claim 1, which is characterized in that the signal acquisition circuit is also in this
When the intensity of voice signal is more than the critical value, which is captured to generate another sound clip signal, this is another
One sound clip signal and the corresponding different sound clip of the sound clip signal, the decision circuitry are one first decision circuitry,
And first decision circuitry is also judging whether the corresponding sound clip of another sound clip signal is vagitus, and
The vagitus detection circuit also includes：
One second decision circuitry is coupled to first decision circuitry, to determine that the voice signal is according to those judging results
No is vagitus.
13. a kind of vagitus detection method, including：
When the intensity of a voice signal is more than a critical value, which is captured to generate sound clip letter
Number, wherein the time span of the corresponding sound clip of the sound clip signal is within a particular range；
Capture multiple characteristic values of the sound clip signal；And
According to those characteristic values come whether judge the sound clip corresponding to the sound clip signal be vagitus.
14. vagitus detection method as claimed in claim 13, which is characterized in that captured the voice signal to produce
The step of raw sound clip signal includes：
When the intensity of the voice signal is more than the critical value, start to capture the voice signal, until the voice signal
Intensity have arrived at the upper limit of the particular range less than the critical value or an acquisition time, to generate sound clip letter
Number.
15. vagitus detection method as claimed in claim 14, which is characterized in that captured the voice signal to produce
The step of raw sound clip signal also includes：
If because when the acquisition time has arrived at the upper limit of the particular range and generates the sound clip signal, from when the acquisition
Between reach the particular range time point of the upper limit start to capture the voice signal to generate next sound clip letter
Number.
16. vagitus detection method as claimed in claim 13, it is characterised in that the particular range be 0.5 second to 3 seconds it
Between.
17. vagitus detection method as claimed in claim 13, which is characterized in that also include：
Believed to generate sound after sampling frequency conversion to be sampled to the voice signal using a fixed sampling frequency
Number；
Voice signal carries out noise Processing for removing to generate voice signal after a noise is eliminated after being converted to the sampling frequency；And
Voice signal carries out Gain tuning processing after being eliminated to the noise, to generate voice signal after the pretreatment；
The step of wherein being captured to the voice signal to generate the sound clip signal is to voice signal after the pretreatment
It is captured to generate the sound clip signal.
18. vagitus detection method as claimed in claim 13, which is characterized in that capture those of the sound clip signal
The step of characteristic value includes：
Multiple sound frames are taken out from the sound clip signal；
Fourier transform is carried out to generate sound frame after multiple Fourier transforms to those sound frames；
Sound frame after those Fourier transforms is filtered, to generate sound frame after multiple filtering；
Discrete cosine transform is carried out to generate the multiple characteristic parameters for corresponding to each sound frame to sound frame after those filtering；And
Those characteristic values of the sound clip signal are generated according to corresponding to those characteristic parameters of each sound frame.
19. vagitus detection method as claimed in claim 18, which is characterized in that capture those of the sound clip signal
The step of characteristic value also includes：
Those sound frames are handled to generate multiple window function sound frames according to a window function；
Wherein carry out Fourier transform to those sound frames includes the step of sound frame after multiple Fourier transforms to generate：
Fourier transform is carried out to generate sound frame after those Fourier transforms to those window function sound frames.
20. vagitus detection method as claimed in claim 18, which is characterized in that capture those of the sound clip signal
The step of characteristic value also includes：
High-pass filtering operation is carried out to the sound clip, to generate a pre-emphasis signal；
The step of wherein taking out those sound frames from the sound clip signal include：
Those sound frames are taken out from the pre-emphasis signal.
21. vagitus detection method as claimed in claim 13, which is characterized in that capture those of the sound clip signal
The step of characteristic value includes：
Multiple sound frames are taken out from the sound clip signal, wherein those characteristic values correspond to those sound frames respectively；
Wherein according to those characteristic values come judge the corresponding sound clip of the sound clip signal whether be vagitus step
Suddenly include：
Judge that the sound clip is believed according to the quantity of multiple medians of those characteristic values, multiple interquartile ranges and those sound frames
Whether number corresponding sound clip is vagitus.
22. vagitus detection method as claimed in claim 13, which is characterized in that judge the sound according to those characteristic values
The step of whether tablet section is vagitus include：
It is to judge the corresponding sound clip of the sound clip signal according to those characteristic values to machine algorithm using one
No is vagitus.
23. vagitus detection method as claimed in claim 22, which is characterized in that the Zhi Xiangji algorithms are with radial direction
The Zhi Xiangji algorithms of basis function core.
24. vagitus detection method as claimed in claim 13, which is characterized in that also include：
When the intensity of the voice signal is more than the critical value, which is captured to generate another sound clip letter
Number, wherein another sound clip signal and the corresponding different sound clip of the sound clip signal；
Judge whether another sound clip signal is vagitus；And
The voice signal is determined according to the vagitus judging result of the sound clip signal and another sound clip signal
Whether it is vagitus.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710044541.8A CN108335704A (en) | 2017-01-19 | 2017-01-19 | Vagitus detection circuit and relevant detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710044541.8A CN108335704A (en) | 2017-01-19 | 2017-01-19 | Vagitus detection circuit and relevant detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108335704A true CN108335704A (en) | 2018-07-27 |
Family
ID=62922984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710044541.8A Pending CN108335704A (en) | 2017-01-19 | 2017-01-19 | Vagitus detection circuit and relevant detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108335704A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111798830A (en) * | 2020-07-07 | 2020-10-20 | 云知声智能科技股份有限公司 | Baby pacifying robot and pacifying method based on baby pacifying robot |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1564245A (en) * | 2004-04-20 | 2005-01-12 | 上海上悦通讯技术有限公司 | Stunt method and device for baby's crying |
| CN2692757Y (en) * | 2004-04-20 | 2005-04-13 | 上海上悦通讯技术有限公司 | Multifunctional baby monitoring teaching apparatus |
| CN1776807A (en) * | 2004-11-15 | 2006-05-24 | 松下电器产业株式会社 | Sound identifying system and safety device having same |
| CN103021406A (en) * | 2012-12-18 | 2013-04-03 | 台州学院 | Robust speech emotion recognition method based on compressive sensing |
| CN103295584A (en) * | 2012-02-24 | 2013-09-11 | 索尼公司 | Voice data detecting device, voice monitoring system and method thereof |
| CN105021275A (en) * | 2015-07-15 | 2015-11-04 | 吉林大学 | Time domain dynamic characteristic analysis-based intra-vehicle sound quality prediction method |
-
2017
- 2017-01-19 CN CN201710044541.8A patent/CN108335704A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1564245A (en) * | 2004-04-20 | 2005-01-12 | 上海上悦通讯技术有限公司 | Stunt method and device for baby's crying |
| CN2692757Y (en) * | 2004-04-20 | 2005-04-13 | 上海上悦通讯技术有限公司 | Multifunctional baby monitoring teaching apparatus |
| CN1776807A (en) * | 2004-11-15 | 2006-05-24 | 松下电器产业株式会社 | Sound identifying system and safety device having same |
| CN103295584A (en) * | 2012-02-24 | 2013-09-11 | 索尼公司 | Voice data detecting device, voice monitoring system and method thereof |
| CN103021406A (en) * | 2012-12-18 | 2013-04-03 | 台州学院 | Robust speech emotion recognition method based on compressive sensing |
| CN105021275A (en) * | 2015-07-15 | 2015-11-04 | 吉林大学 | Time domain dynamic characteristic analysis-based intra-vehicle sound quality prediction method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111798830A (en) * | 2020-07-07 | 2020-10-20 | 云知声智能科技股份有限公司 | Baby pacifying robot and pacifying method based on baby pacifying robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102318325B (en) | Controlling an adaptation of a behavior of an audio device to a current acoustic environmental condition | |
| CN101600144B (en) | Method and system for obtaining a plurality of parameters of electro-acoustic product by adopting continuous logarithmic swept-frequency signal | |
| KR20030076560A (en) | Method and apparatus for removing noise from electronic signals | |
| US20100329471A1 (en) | Ambient noise compensation system | |
| EP3671743A1 (en) | Voice activity detection method | |
| CN111147079A (en) | Data acquisition method and device with adaptive and adjustable sampling frequency | |
| CN104244127A (en) | Heart rate detection method applied to ear phone and ear phone capable of detecting heart rate | |
| KR102429152B1 (en) | Deep learning voice extraction and noise reduction method by fusion of bone vibration sensor and microphone signal | |
| TWI597720B (en) | Baby cry detection circuit and associated detection method | |
| CN106504760B (en) | Broadband ambient noise and speech Separation detection system and method | |
| CN113674763B (en) | Method, system, device and storage medium for identifying whistle by utilizing line spectrum characteristics | |
| CN112033656A (en) | Mechanical system fault detection method based on broadband spectrum processing | |
| CN108040315A (en) | A kind of test machine of computer-readable recording medium and the application medium | |
| CN108335704A (en) | Vagitus detection circuit and relevant detection method | |
| CN114974301A (en) | Abnormal sound detection method, computer readable storage medium and electronic device | |
| CN102988041A (en) | Selective cardiac-magnetic signal averaging method in signal noise suppression | |
| Ruan et al. | Use AF-CNN for end-to-end fiber vibration signal recognition | |
| CN114280368A (en) | Burst signal detection method and detection system in complex environment | |
| CN106714067A (en) | Automatic detection method on production line and automatic detection device thereof | |
| CN113599052A (en) | Snore monitoring method and system based on deep learning algorithm and corresponding electric bed control method and system | |
| CN102759571B (en) | Product quality test process and test device | |
| CN108344501A (en) | Resonance identification and removing method and device in a kind of application of signal correlation | |
| CN110556128B (en) | Voice activity detection method and device and computer readable storage medium | |
| CN105989834A (en) | Voice recognition apparatus and voice recognition method | |
| CN107221338A (en) | Sound wave extraction element and extracting method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20181224 Address after: 2 on the 10th floor, 1 Taiyuan Second Street, Zhubei City, Hsinchu County, Taiwan, China Applicant after: Juxing Electronics Co.,Ltd. Address before: 1/2, 4th floor, 26 Taiyuan Street, Zhubei City, Hsinchu County, Taiwan, China Applicant before: MSTAR SEMICONDUCTOR Inc. |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20190725 Address after: Area 1501, Innovation Building A, Software Park, Xiamen Torch High-tech Zone, Xiamen City, Fujian Province Applicant after: Xiamen Xingchen Technology Co.,Ltd. Address before: 2 on the 10th floor, 1 Taiyuan Second Street, Zhubei City, Hsinchu County, Taiwan, China Applicant before: Juxing Electronics Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| CB02 | Change of applicant information |
Address after: 361006 1501, zone a, innovation building, software park, Xiamen Torch High tech Zone, Xiamen, Fujian Applicant after: Xiamen Xingchen Technology Co.,Ltd. Address before: 361006 1501, zone a, innovation building, software park, Xiamen Torch High tech Zone, Xiamen, Fujian Applicant before: Xiamen Xingchen Technology Co.,Ltd. Address after: 361006 1501, zone a, innovation building, software park, Xiamen Torch High tech Zone, Xiamen, Fujian Applicant after: Xingchen Technology Co.,Ltd. Address before: 361006 1501, zone a, innovation building, software park, Xiamen Torch High tech Zone, Xiamen, Fujian Applicant before: Xiamen Xingchen Technology Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180727 |
|
| RJ01 | Rejection of invention patent application after publication |