CN106529433B - Queue march in step degree evaluation method based on voice signal - Google Patents
Queue march in step degree evaluation method based on voice signal Download PDFInfo
- Publication number
- CN106529433B CN106529433B CN201610937669.2A CN201610937669A CN106529433B CN 106529433 B CN106529433 B CN 106529433B CN 201610937669 A CN201610937669 A CN 201610937669A CN 106529433 B CN106529433 B CN 106529433B
- Authority
- CN
- China
- Prior art keywords
- curve
- voice signal
- signal
- queue
- paces
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention is by being analyzed and processed queue step voice signal of reviewing troops, for propose by queue step voice signal analyze dequeue paces neat degree, provide a kind of completely new digitlization, Assessment mode for the uniformity evaluation for queue of reviewing troops.The technical solution adopted by the present invention is that, queue march in step degree evaluation method based on voice signal, steps are as follows: 1) acquisition of sound and noise reduction 2) extractions of paces by the extraction to paces coordinate, determines specific location of each paces in voice signal；3) S curve that the analysis of march in step degree passes through the building neat paces voice signal of ideal standard, the cumlative energy curve of itself and actual sound signal is set to carry out correlation analysis, related coefficient is bigger, illustrate that practical step signal is more close with the neat step voice signal of standard, i.e., practical step signal is more neat.Present invention is mainly applied to step voice signals to be analyzed and processed.
Description
Technical field
The present invention evaluates the march in step journey for queue of reviewing troops by being analyzed and processed to queue step voice signal of reviewing troops
Degree, to make corresponding judgement to the consistency of team posture.Embody formation, the standard of queue inspection and achievement more
Add intuitive and scientific.Concretely relate to the queue march in step degree evaluation method based on voice signal.
Background technique
No matter formation is all the exercise sense of organizational discipline, steel one's will product in colleges and universities' military training or in military training
The important way of matter is to be generally accepted team training mode in China.And the neat degree of queue paces is even more queue instruction
Practice the concentrated reflection of effect.Currently, queue march in step degree is directed to, still without a kind of scientific and effective evaluation method.Especially
It is the evaluation of effect of reviewing troops for queue, is all made of in such a way that reviewer's subjective feeling is evaluated substantially.It is instructed in queue
Guidance also not scientific, fully relies on mechanization repetition training to achieve the desired results during practicing.Digitlization with
Today of information-based high speed development, this inevitably seems relatively backward.For existing technical thought, it can be used to carry out queue
The Digital evaluation mode of uniformity detection mainly has moving object based on computer vision to identify and track, and based on biography
The movement statistics of sensor network compares such two kinds of Research Thinkings.
Computer vision is to obtain and handle the technology of information by the vision mechanism of computer mould personification.It is mainly
The image of human motion is recorded, and detects, extract from image, tracking people and a series of actions of people is analyzed and retouched
It states.Solution based on sensor network is identified by acquiring and handling the acceleration signal generated when human motion
Different human actions identifies position of human body by acquisition ground vibration signal.Then its otherness and correlation are calculated.
Time cost required for both the above method is developed, resources costs are all relatively high, and difficulty is larger.So this hair
It is bright to use the thinking analyzed by queue voice signal.The current research based on voice signal both at home and abroad is mainly reflected in language
The fields such as sound identification and processing.The 1960s, foreign countries formd the theory and calculation of a series of Digital Signal Processing of maturations
Method, such as digital filter.This provides the research in terms of for sound uniformity to a large amount of reference.Meanwhile there are also permitted both at home and abroad
The research of automobile model identification, animal species judgement, mechanical disorder detection mostly based on voice signal etc..But these grind
The popular feature studied carefully is all based on frequency domain and analyzes signal.And for the step voice signal for queue of reviewing troops, neat person and
The footsteps that irregular person issues is almost the same on frequency domain, it is difficult to distinguish, it is therefore desirable to from time domain angle to signal into
Row analysis.
In modern medical service clinical diagnosis, there is a kind of clinical diagnosis technology based on analysis of PCG Signal.This technology benefit
Time-frequency combination analysis is carried out to heartbeat sound with time-frequency distributions model of wavelet transformation etc..This is believed for the present invention by sound
Number to the neat degree of queue paces carry out evaluation provide inspiration.
Bibliography:
The Nanjing Detection And Tracking [D] of personnel step signal of the Kong Fei based on multichannel sensor: Institutes Of Technology Of Nanjing,
2006.
The dynamic hand gesture recognition of Zhang Guoliang, Wang Zhanni, the Appliance computer vision such as king field summarize [J] Huaqiao University journal
(natural science edition), 2014, (6): 653-658.
The detection of the cardiechema signals such as Guo Juntao, Lin Sencai, Tian Xiaodong and processing [J] modern instrument, 2008,14 (2):
36-38.
Design method [J] the scientific and technological information of Li Weizhi based on Underwater Acoustic channels: scholarly edition, 2008, (17).
Summary of the invention
In order to overcome the deficiencies of the prior art, the present invention is directed to propose being obtained by carrying out analysis to queue step voice signal
The neat degree of queue paces provides the completely new digitlization of one kind, Assessment mode for the uniformity evaluation for queue of reviewing troops.
The technical solution adopted by the present invention is that the queue march in step degree evaluation method based on voice signal, steps are as follows:
1) acquisition of sound and noise reduction
Firstly, from hardware point of view, using the sound pick-up with directive property, and according to the propagation characteristic section of live sound
The placement angle for choosing sound pick-up and position are learned, the step voice signal of high s/n ratio as far as possible is acquired with this；Using time-frequency knot
The nonlinear filtering noise reduction based on wavelet transformation closed；
2) extraction of paces
Zero-crossing rate processing is carried out to collected voice signal first, zero-crossing rate ZCR (zero-crossing rate) is
Refer to the ratio of the sign change of a signal, the formula of zero-crossing rate are as follows:
Wherein, ZnFor n-th point of zero-crossing rate, N-1 is that zero passage window is long, XnIt (m) is m-th point after nth point of amplitude；
Sgn [X] is sign function, and whether the unknown quantity X for judging in bracket is greater than 0, it may be assumed that
By the extraction to paces coordinate, specific location of each paces in voice signal is determined；
3) march in step degree is analyzed
By constructing the S curve of the neat paces voice signal of ideal standard, make the cumlative energy of itself and actual sound signal
Curve carries out correlation analysis, and related coefficient is bigger, illustrates that practical step signal is more close with the neat step voice signal of standard,
I.e. practical step signal is more neat.
The construction step of S curve is the fundamental formular and relevant parameter for needing to obtain logic distribution, distribution function F (x)
With the fundamental formular of probability density f (x) are as follows:
Wherein
Above-mentioned distribution function F (x) is thus referred to as population growth curve, i.e. S curve, when independent variable x is represented existing for population
Between sequence, s is slope of the curve at μ value, and μ value is S curve maximum slope corresponding time point, and μ value appears in energy sum
At 1/2.
Calculate the cumlative energy of practical step signal are as follows:
Wherein E (i) is signal corresponding energy at i-th；
According to the theoretical basis of logic distribution, if the integral distribution curve of its corresponding ideal s curve are as follows:
Wherein L be actual waveform cumlative energy and, i represents i-th of time point, and μ is that actual waveform cumlative energy reaches
Time point when L/2, s are the key that entire formula, and system is by acquiring the s of the point of actual waveform cumlative energy 40% to 60%
It is averaged after value, uses the average value approximation that acquires as the s value of ideal S curve.
Using Pearson correlation coefficients (Pearson correlation coefficient) to two curve S1And S2Into
Row correlation test, Pearson correlation coefficients calculation formula are as follows:
Wherein ρX, YThe related coefficient of two variables of X and Y is represented, molecule cov (X, Y) is the covariance of X and Y, denominator σ X σ Y
It is the product of two variable standard deviations, E is expectation, it is clear that linear when two variables it is required that the standard deviation of X and Y all cannot be 0
When relationship enhances, related coefficient tends to 1 or -1.Tend to 1 when positive correlation, tend to -1 when negatively correlated, when two variable independence phases
Relationship number is 0, according to correlation test principle, takes X and Y is respectively the cumlative energy distribution waveform array of original sound signal
With corresponding ideal S curve array, the related coefficient calculated can represent neat degree.
Wavelet transformation is boir4_4 small echo.
The features of the present invention and beneficial effect are:
The present invention provides the evaluation side of the completely new digitlization of one kind, automation for the evaluation of queue march in step degree
Formula.It avoids because of the problem that evaluation result caused by the human factors such as individual cognition difference is not allowed.It can greatly increase simultaneously
Drill and review troops scene science and interest.The method can be widely applied to army or colleges and universities formation and
It reviews troops in joint performance.
Detailed description of the invention:
Fig. 1 original sound signal.
Signal after Fig. 2 wavelet de-noising.
Fig. 3 zero-crossing rate.
The energy accumulation distribution curve of mono- step voice signal of Fig. 4.
Fig. 5 population growth s curve.
Fig. 6 logic distribution probability density curve and integral distribution curve.
The cumlative energy curve of Fig. 7 actual waveform.
Fig. 8 approximate ideal S curve.
Fig. 9 invention flow chart.
Specific embodiment
It is obvious that step sound of the neat queue when advancing be also it is neat, concentrating, rhythmical, foot landing and
The hanging time can obviously be distinguished.And the queue that uniformity is relatively low, the step sound in traveling process be then it is mixed and disorderly,
It is dispersion, rhythmic lower, and the time of foot landing is that dispersion is irregular.It therefore, can by further analysing and comparing
It is obviously distinguished with the neat degree to queue step sound.
This programme will be made of following steps: point of the acquisition of sound and noise reduction, the extraction of paces and uniformity
Analysis.
1. acquisition and the noise reduction of sound
The premise analyzed queue step sound is to obtain clear pure step voice signal.Firstly, from hardware
Angle is set out, and chooses sound pick-up using the sound pick-up having compared with high directivity, and according to the propagation characteristic science of live sound
Placement angle and position acquire the step voice signal of high s/n ratio as far as possible with this.Fig. 1 is the step sound that actual acquisition arrives
Then signal, that is, original sound signal further carries out noise reduction from software respective to it.Due to queue review troops scene sound at
Divide complexity, other than echo signal, that is, step voice signal, there is also musical sound, personnel's sound of speech and other various complexity
Natural noise.And noise signal and step signal have a big chunk to overlap on frequency domain, and it is traditional based on impact
The noise reduction mode effect of response filter is very unsatisfactory.
The nonlinear filtering noise reduction mode based on wavelet transformation that the present invention uses a kind of time-frequency to combine.Wavelet de-noising mode
It has been obtained in the time-frequency binding analysis of frequency overlapping commonly used.According to wavelet theory, for unstable letter
Number, the shake of signal can be effectively filtered out using wavelet analysis.Pass through the time subdivision to high-frequency region, the frequency of low frequency region
Rate subdivision, system can interference component and key signaling components in effective resoluting signal, keep next step operation more accurate
Property.It is tested by great amount of samples, we have selected boir4_4 small echo, and this small echo is biorthogonal compactly supported wavelet, have linear
Phase, and had mature application in JPEG2000 standard.Meanwhile this wavelet shapes and time domain voice signal have centainly
Degree of fitting can be realized preferable noise reduction effect.Fig. 2 is signal of the actual signal after wavelet de-noising, it can be seen that scene
Noise ingredient is obviously filtered out.
2. the extraction of paces
It needs to extract paces information after carrying out noise reduction process to signal, to carry out independent analysis to each step.
The present invention carries out zero-crossing rate processing to collected voice signal first, and zero-crossing rate (zero-crossing rate, ZCR) refers to
The ratio of the sign change of one signal, such as signal become negative or reversed from positive number.This feature is in Speech comparison, voice
Identification and music information retrieval field are used widely, and are the main features classified to percussion sound.It is usually used in analyzing
The rhythmic feature of one section of sound.The formula of zero-crossing rate are as follows:
Wherein, ZnFor n-th point of zero-crossing rate, N-1 is that zero passage window is long, XnIt (m) is m-th point after nth point of amplitude.
Sgn [X] is sign function, and whether the unknown quantity X for judging in bracket is greater than 0, it may be assumed that
By the calculating of zero-crossing rate, can by multiple paces interval and single paces separate.I.e. in compartment, mistake
The value of zero rate is always zero；In the single paces stage, comparatively dense is accordingly compared in the distribution of zero-crossing rate.If Fig. 3 is after zero-crossing rate is handled
Sound waveform.Pass through the extraction to paces coordinate, it can determine specific location of each paces in voice signal.
3. march in step degree is analyzed
Paces need individually to carry out uniformity analysis to each step after the completion of extracting.Such as Fig. 1, in step sound signal waveform
In, where can obviously distinguish is the sound for stepping a step, where is interval between two steps, this is because in wave-shape amplitude
Difference, that is, the difference in intensity of sound.When ignoring factors and everyone footsteps such as sound reflection, sound superposition
When the difference of sound, can approximately think the size of intensity of sound be proportional to generate this step sound number number.Time domain
On signal waveform in, the intensity distribution of sound can represent the number that different moments foot when stepping same step lands, that is,
This queue member steps the neat degree of this step.Based on assumed above, the intensity distribution of step voice signal is divided
Analysis.
It takes out a step therein and does cumlative energy distribution curve, obtain curve as shown in Figure 4, analyze and send out by mass data
It is existing, the curve tendency approximate Logic profile accumulation energy curve.Generally existing population growth curve namely in nature.?
In natural environment, the growth curve of population is " S " type curve (also referred to as logistic curve), such as Fig. 5.Population reaches ring
The maximum value that border can be born, the referred to as laden weight or load capacity of environment are indicated with " K ".Population quantity when reaching K/2,
Population quantity linearly rises, this period claims exponential phase of growth, and this point is to influence the key point of population quantity.If
The problem of analyzing step sound with the thinking of logic distribution S curve, it is found that ignore the characteristics such as sound transmission, diverging, reflection
Influence, the cumlative energy for the step voice signal that a step is neatly concentrated is gradually increasing at the beginning, connects to foot and ground
When the key point of touching, acoustic energy growth rate highest, as foot leaves ground, acoustic energy growth rate is reduced to zero, accumulates energy
Amount reaches maximum value.Therefore, step voice signal is similar S curve in Energy distribution.
Therefore itself and actual sound signal can be made by the S curve of the building neat paces voice signal of ideal standard
Cumlative energy curve carries out correlation analysis, and related coefficient is bigger, illustrates practical step signal and the neat footsteps message of standard
Number more close, i.e., practical step signal is more neat.Therefore the degree of relevancy of the two just embodies paces to a certain extent
Neat degree.
To construct ideal S curve, need to obtain the fundamental formular and relevant parameter of logic distribution.Logic is distributed
It is the mathematical model of a comparative maturity, the fundamental formular of distribution function F (x) and probability density f (x) are as follows:
Wherein
Above-mentioned distribution function F (x) is thus referred to as population growth curve, i.e. S curve, when independent variable x is represented existing for population
Between sequence, s is slope of the curve at μ value, and μ value is S curve maximum slope corresponding time point, and μ value appears in energy sum
At 1/2.
Fig. 6 is corresponding probability density curve and integral distribution curve.
Calculate the cumlative energy of practical step signal are as follows:
Wherein E (i) is signal corresponding energy at the time i-th.
According to the theoretical basis of logic distribution, if the integral distribution curve of its corresponding ideal s curve are as follows:
Wherein L be actual waveform cumlative energy and, i represents i-th of time point, and μ is that actual waveform cumlative energy reaches
Time point when L/2, s are the key that entire formula, and system is by acquiring the s of the point of actual waveform cumlative energy 40% to 60%
It is averaged after value, uses the average value approximation that acquires as the s value of ideal S curve.One column voice signal is carried out by the method
It calculates, the cumlative energy curve and approximate ideal S curve of available actual waveform, such as the figure that Fig. 7, Fig. 8 are wherein two steps
Picture.
Finally, the ideal S curve corresponding to the integrated energy curve of actual waveform carries out correlation test.It is corresponding
Related coefficient can represent the neat degree of the step voice signal.
Common Pearson correlation coefficients (Pearson correlation coefficient) carries out correlation test.Skin
Ademilson related coefficient is also Pearson's Coefficient of production-moment correlation, is the statistic for reacting two variable similarity degrees.In other words
The similarity of two vectors can be used to calculate.Pearson correlation coefficients calculation formula is as follows:
Wherein ρX, YThe related coefficient of two variables of X and Y is represented, molecule cov (X, Y) is the covariance of X and Y, denominator σ X σ Y
It is the product of two variable standard deviations, E is expectation.Obviously the standard deviation for requiring X and Y all cannot be 0.It is linear when two variables
When relationship enhances, related coefficient tends to 1 or -1.Tend to 1 when positive correlation, tends to -1 when negatively correlated.When two variable independence phases
Relationship number is 0.According to correlation test principle, takes X and Y is respectively the cumlative energy distribution waveform array of original sound signal
With corresponding ideal S curve array, the related coefficient calculated can represent neat degree.
Realize that flow chart of the invention is integrally as shown in Figure 9.According to the thinking of technical solution, first from hardware point of view, lead to
The professional recording microphone for crossing heart-shaped direction, is directed at the step audio direction for queue of reviewing troops, collects the sound of high s/n ratio as far as possible
Sound signal.Next, can be analyzed and processed by softwares such as labview or matlab to collected voice signal.Letter
Number processing is mainly made of following steps: the noise reduction of original signal, the extraction of single paces, the uniformity analysis of paces.
The neat degree of collected step voice signal can be obtained after carrying out respective handling to signal according to above-mentioned theory basis.
Claims (3)
1. a kind of queue march in step degree evaluation method based on voice signal, characterized in that steps are as follows:
1) acquisition of sound and noise reduction
Firstly, being selected from hardware point of view using the sound pick-up with directive property, and according to the propagation characteristic science of live sound
Placement angle and the position of sound pick-up are taken, the step voice signal of high s/n ratio as far as possible is acquired with this；It is combined using time-frequency
Nonlinear filtering noise reduction based on wavelet transformation；
2) extraction of paces
Zero-crossing rate processing is carried out to collected voice signal first, zero-crossing rate ZCR (zero-crossing rate) refers to one
The ratio of the sign change of a signal, the formula of zero-crossing rate are as follows:
Wherein, ZnFor n-th point of zero-crossing rate, N-1 is that zero passage window is long, XnIt (m) is m-th point after nth point of amplitude；
Sgn [X] is sign function, and whether the unknown quantity X for judging in bracket is greater than 0, it may be assumed that
By the extraction to paces coordinate, specific location of each paces in voice signal is determined；
3) march in step degree is analyzed
By constructing the S curve of the neat paces voice signal of ideal standard, make its cumlative energy curve with actual sound signal
Correlation analysis is carried out, related coefficient is bigger, illustrates that practical step signal is more close with the neat step voice signal of standard, i.e., in fact
Border step signal is more neat；
The construction step of S curve is the fundamental formular and relevant parameter for needing to obtain logic distribution, distribution function F (x) and general
The fundamental formular of rate density f (x) are as follows:
Wherein
Above-mentioned distribution function F (x) is thus referred to as population growth curve, i.e. S curve, and independent variable x represents time sequence existing for population
Column, s are slope of the curve at μ value, and μ value is S curve maximum slope corresponding time point, and μ value appears in the 1/2 of energy sum
Place；
Calculate the cumlative energy of practical step signal are as follows:
Wherein E (i) is signal corresponding energy at i-th；
According to the theoretical basis of logic distribution, if the integral distribution curve of its corresponding ideal s curve are as follows:
Wherein L be actual waveform cumlative energy and, i represents i-th of time point, when μ is that actual waveform cumlative energy reaches L/2
Time point, s is the key that entire formula, after s value of the system by the point for acquiring actual waveform cumlative energy 40% to 60%
It is averaged, uses the average value approximation that acquires as the s value of ideal S curve.
2. the queue march in step degree evaluation method based on voice signal as described in claim 1, characterized in that use Pierre
Gloomy related coefficient (Pearson correlation coefficient) is to two curve S1And S2Carry out correlation test, Pierre
Gloomy related coefficient calculation formula is as follows:
Wherein ρX, YThe related coefficient of two variables of X and Y is represented, molecule cov (X, Y) is the covariance of X and Y, and denominator σ X σ Y is two
The product of a variable standard deviation, E are expectations, it is clear that it is required that the standard deviation of X and Y all cannot be 0, when the linear relationship of two variables
When enhancing, related coefficient tends to 1 or -1, and when positive correlation tends to 1, tends to -1 when negatively correlated, the phase relation when two variable independence
Number is 0, according to correlation test principle, takes X and Y is respectively the cumlative energy distribution waveform array of original sound signal and right
Ideal S curve array is answered, the related coefficient calculated can represent neat degree.
3. the queue march in step degree evaluation method based on voice signal as described in claim 1, characterized in that wavelet transformation
For boir4_4 small echo.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610937669.2A CN106529433B (en) | 2016-10-25 | 2016-10-25 | Queue march in step degree evaluation method based on voice signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610937669.2A CN106529433B (en) | 2016-10-25 | 2016-10-25 | Queue march in step degree evaluation method based on voice signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106529433A CN106529433A (en) | 2017-03-22 |
| CN106529433B true CN106529433B (en) | 2019-07-16 |
Family
ID=58293282
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610937669.2A Expired - Fee Related CN106529433B (en) | 2016-10-25 | 2016-10-25 | Queue march in step degree evaluation method based on voice signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106529433B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107688787B (en) * | 2017-09-01 | 2020-09-29 | 宜宾学院 | Near-end interphalangeal joint line identification method based on Gabor wavelet |
| CN110392301B (en) * | 2019-07-05 | 2022-09-23 | 湖北盟道信息科技有限公司 | Method and system for automatically muting audio of multiple student terminals |
| CN112529473B (en) * | 2020-12-28 | 2021-12-24 | 河北中兴汽车制造有限公司 | Evaluation method for simulating working condition validity of fuel vehicle by electric learner-driven vehicle |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101632587A (en) * | 2009-08-05 | 2010-01-27 | 南京大学 | Tread signal extracting method based on wavelet transformation |
| CN102163427A (en) * | 2010-12-20 | 2011-08-24 | 北京邮电大学 | Method for detecting audio exceptional event based on environmental model |
| CN102664006A (en) * | 2012-04-14 | 2012-09-12 | 中国人民解放军国防科学技术大学 | Abnormal voice detecting method based on time-domain and frequency-domain analysis |
| CN104409078A (en) * | 2014-12-11 | 2015-03-11 | 黑龙江真美广播通讯器材有限公司 | Abnormal noise detection and recognition system |
| CN104464726A (en) * | 2014-12-30 | 2015-03-25 | 北京奇艺世纪科技有限公司 | Method and device for determining similar audios |
| CN104637481A (en) * | 2015-01-22 | 2015-05-20 | 沈阳大学 | Vehicle model audio feature extracting method based on LMD (local mean decomposition) and energy projection methods |
| CN105845144A (en) * | 2016-03-21 | 2016-08-10 | 陈宁 | Intelligent health management system for realizing animal sound and form translation function |
-
2016
- 2016-10-25 CN CN201610937669.2A patent/CN106529433B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101632587A (en) * | 2009-08-05 | 2010-01-27 | 南京大学 | Tread signal extracting method based on wavelet transformation |
| CN102163427A (en) * | 2010-12-20 | 2011-08-24 | 北京邮电大学 | Method for detecting audio exceptional event based on environmental model |
| CN102664006A (en) * | 2012-04-14 | 2012-09-12 | 中国人民解放军国防科学技术大学 | Abnormal voice detecting method based on time-domain and frequency-domain analysis |
| CN104409078A (en) * | 2014-12-11 | 2015-03-11 | 黑龙江真美广播通讯器材有限公司 | Abnormal noise detection and recognition system |
| CN104464726A (en) * | 2014-12-30 | 2015-03-25 | 北京奇艺世纪科技有限公司 | Method and device for determining similar audios |
| CN104637481A (en) * | 2015-01-22 | 2015-05-20 | 沈阳大学 | Vehicle model audio feature extracting method based on LMD (local mean decomposition) and energy projection methods |
| CN105845144A (en) * | 2016-03-21 | 2016-08-10 | 陈宁 | Intelligent health management system for realizing animal sound and form translation function |
Non-Patent Citations (3)
| Title |
|---|
| 基于多通道传感器的人员脚步信号的探测与跟踪;孔飞;《中国优秀博硕士学位论文全文数据库（硕士）信息科技辑》;20070115;第I140-335页 |
| 基于局部频谱特征的脚步声识别算法研究;余瑶;《中国优秀硕士学位论文全文数据库信息科技辑》;20150215;第I136-222页 |
| 基于脚步声身份识别的算法研究;张瑞兴;《中国优秀硕士学位论文全文数据库信息科技辑》;20150215;第I138-1097页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106529433A (en) | 2017-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108962237B (en) | Hybrid speech recognition method, device and computer readable storage medium | |
| Han et al. | Convolutional Neural Networks with Binaural Representations and Background Subtraction for Acoustic Scene Classification. | |
| Kao et al. | A human-machine interaction technique: hand gesture recognition based on hidden Markov models with trajectory of hand motion | |
| US20170143226A1 (en) | Action Recognition Method and Device Based on Surface Electromyography Signal | |
| CN106529433B (en) | Queue march in step degree evaluation method based on voice signal | |
| CN112259120B (en) | Single-channel human voice and background voice separation method based on convolution cyclic neural network | |
| CN103092971B (en) | A kind of sorting technique for brain-computer interface | |
| CN104795064A (en) | Recognition method for sound event under scene of low signal to noise ratio | |
| CN110069199A (en) | A kind of skin-type finger gesture recognition methods based on smartwatch | |
| CN105448291A (en) | Parkinsonism detection method and detection system based on voice | |
| CN108567418A (en) | A kind of pulse signal inferior health detection method and detecting system based on PCANet | |
| Shahid et al. | A new spike detection algorithm for extracellular neural recordings | |
| CN108958474A (en) | A kind of action recognition multi-sensor data fusion method based on Error weight | |
| CN113539294A (en) | Method for collecting and identifying sound of abnormal state of live pig | |
| CN113116361A (en) | Sleep staging method based on single-lead electroencephalogram | |
| Dang et al. | Acoustic scene classification using convolutional neural networks and multi-scale multi-feature extraction | |
| CN102509548B (en) | Audio indexing method based on multi-distance sound sensor | |
| Jena et al. | Gender recognition of speech signal using knn and svm | |
| Nishizaki et al. | Signal classification using deep learning | |
| CN109730700A (en) | A kind of user emotion based reminding method | |
| Jiang et al. | Speech emotion recognition method based on improved long short-term memory networks | |
| CN109616099A (en) | A kind of men and women's voice gender identification method based on fundamental frequency and smoothing processing | |
| Abel et al. | A data driven approach to audiovisual speech mapping | |
| CN113488069B (en) | Speech high-dimensional characteristic rapid extraction method and device based on generation type countermeasure network | |
| CN115500790A (en) | Method for predicting and analyzing epileptic seizure based on BI-LSTM network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190716 Termination date: 20211025 |