CN108417221B - Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering - Google Patents
Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering Download PDFInfo
- Publication number
- CN108417221B CN108417221B CN201810074257.XA CN201810074257A CN108417221B CN 108417221 B CN108417221 B CN 108417221B CN 201810074257 A CN201810074257 A CN 201810074257A CN 108417221 B CN108417221 B CN 108417221B
- Authority
- CN
- China
- Prior art keywords
- dimensional
- signal
- sound code
- code signal
- recombination
- 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.)
- Active
Links
- 230000004927 fusion Effects 0.000 title claims abstract description 46
- 238000001914 filtration Methods 0.000 title claims abstract description 45
- 238000001514 detection method Methods 0.000 title claims abstract description 30
- 230000006798 recombination Effects 0.000 title claims abstract description 28
- 238000005215 recombination Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 14
- 238000013145 classification model Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 9
- 238000012706 support-vector machine Methods 0.000 claims description 17
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
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
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/16—Vocoder architecture
-
- 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
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
-
- 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
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/26—Pre-filtering or post-filtering
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Complex Calculations (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention discloses a digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering, which comprises the following steps of: performing two-dimensional signal recombination on the one-dimensional acoustic code signal, performing fusion filtering processing on the recombined two-dimensional acoustic code signal, and performing model training on the two-dimensional acoustic code signal subjected to fusion filtering according to features obtained by attribute reduction to obtain an SVM classification model; and carrying out two-dimensional signal recombination and fusion filtering processing on the digital interphone sound code signal sample to be detected, using an SVM classification model to test and detect the characteristics of the two-dimensional signal subjected to the fusion filtering processing and obtained according to attribute reduction, and carrying out decision level fusion on the test and detection result of each two-dimensional signal by adopting a voting method to obtain a final sound code signal type detection result. The invention adopts a signal two-dimensional recombination method to carry out two-dimensional operation on the one-dimensional sound code signal, thereby better extracting the characteristics of the sound code signal and improving the accuracy of sound code signal type detection.
Description
Technical Field
The invention belongs to the field of digital signal processing, and particularly relates to a digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering.
Background
The digital interphone is designed by adopting a digital technology. The digital interphone digitizes the voice signal and transmits the signal in a digital coding mode, and all the modulations on the interphone transmission frequency are digital. Compared with analog interphone, digital interphone has the advantages of strong anti-interference ability, good conversation quality, high frequency utilization rate, good confidentiality, supporting data service, convenient error-free relay, etc., and has wide application. In some specific locations, it is often necessary to detect the type of vocoded digital signals in digital walkie-talkies in order to further identify the content of the digital walkie-talkie transmissions.
The key to the detection of the type of vocoded signals is the feature extraction of the signals. The existing acoustic code signal type detection algorithms are used for carrying out acoustic code type detection on one-dimensional acoustic code signals, and because the one-dimensional acoustic code signals are easily influenced by noise and have obvious intra-class difference, the characteristics of input signals cannot be effectively extracted, and the precision and the stability of the acoustic code signal type detection algorithms need to be improved.
Disclosure of Invention
The invention aims to provide a digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering, which improves the precision and stability of digital interphone sound code type detection and is suitable for high-precision sound code type efficient detection of small sample sound code signals.
The technical scheme for realizing the purpose of the invention is as follows: a digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering comprises the following steps:
step 1, performing two-dimensional signal recombination on a one-dimensional acoustic code signal to obtain a two-dimensional acoustic code signal;
step 2, preprocessing the recombined two-dimensional sound code signal by using a fusion filtering method;
step 3, using a support vector machine as a classifier, and performing model training on the two-dimensional acoustic code signals subjected to fusion filtering according to features obtained by attribute reduction to obtain an SVM classification model;
step 4, carrying out two-dimensional signal recombination and fusion filtering processing on the digital interphone sound code signal sample to be detected;
and 5, testing and detecting the characteristics of the two-dimensional sound code signals subjected to fusion filtering processing according to attribute reduction by using the SVM classification model obtained by training in the step 3, and performing decision-level fusion on the test detection result of each two-dimensional sound code signal by adopting a voting method to obtain a final sound code signal type detection result.
Compared with the prior art, the invention has the remarkable advantages that: (1) the invention adopts a signal two-dimensional recombination method to carry out two-dimension on various one-dimensional sound code signals with obvious internal differences, thereby better extracting the characteristics of the sound code signals and improving the accuracy of sound code signal type detection; (2) the influence of noise is reduced by using a fusion filtering method, and effective characteristics are selected from each group of characteristics of the acoustic code signal by using an attribute reduction method, so that the type detection of the acoustic code signal is more stable and reliable; (3) and decision-level fusion is carried out by a voting method, so that the detection precision is further improved.
Drawings
Fig. 1 is a flow chart of a digital interphone vocoded type detection method based on signal two-dimensional recombination fusion filtering.
Detailed Description
The invention solves the problem of overlarge difference in one-dimensional signals by two-dimensional recombination of signals, and adopts a fusion filtering method to preprocess the sound code signals to remove the influence of noise on the sound code signals. Selecting a plurality of effective characteristics from the preprocessed digital interphone two-dimensional sound code signals by using an attribute reduction method for analysis, inputting the extracted two-dimensional sound code signal characteristics of the sound code signals into an SVM classifier for training to obtain an SVM classification model, and performing class analysis on the test signals subjected to two-dimensional recombination and fusion filtering by using the classification model. And finally, processing the prediction result by using a decision-level fusion method to obtain a label corresponding to each test signal. The method can obtain the high-precision, stable and robust sound code signal detection effect under the condition of smaller training samples.
With reference to fig. 1, a digital interphone vocoding type detection method based on signal two-dimensional recombination fusion filtering includes the following steps:
step 1, performing two-dimensional signal recombination on one-dimensional acoustic code signals, and dividing the one-dimensional acoustic code signals containing D bytes into T rows to obtain two-dimensional acoustic code signals;
step 2, preprocessing the recombined two-dimensional sound code signal by using a fusion filtering method;
step 3, using a support vector machine as a classifier, and performing model training on the two-dimensional acoustic code signals subjected to fusion filtering according to features obtained by attribute reduction to obtain an SVM classification model;
step 4, carrying out two-dimensional signal recombination and fusion filtering processing on the digital interphone sound code signal sample to be detected;
and 5, testing and detecting the characteristics of the two-dimensional sound code signals subjected to fusion filtering processing according to attribute reduction by using the SVM classification model obtained by training in the step 3, and performing decision-level fusion on the test detection result of each two-dimensional sound code signal by adopting a voting method to obtain a final sound code signal type detection result.
Further, step 1 specifically comprises:
assume that there are N different samples: x ═ X1,x2,…,xN|xi∈RD,i∈[1,2,…,N]Where D is the data size of each sample of the one-dimensional vocoded signal, xiIs a sampleN is the number of samples; set of one-dimensional vocoded signal labels Y ═ Y1,y2,…yi…,yN|i=1,2,…,N},yiA label for the ith one-dimensional vocoded signal;
converting the one-dimensional sample information into two-dimensional sample information, where the converted sample may be represented as: x' ═ X11,x12,…,xNT|xij∈RD/T,i∈[1,2,…,N],j∈[1,2,…,T]}
Wherein xijAnd obtaining two-dimensional sample data of the line N x T and the line D/T after the division is finished.
Preferably, the one-dimensional vocoded signal including 189 bytes is divided into 7 rows in step 1.
Further, in the step 2, the fusion filtering is to preprocess the two-dimensional recombined sound code signal by a sound code signal intensity mean value filtering method;
firstly, constructing a 3 x 3 two-dimensional sliding template, replacing the value of one point in a two-dimensional sound code signal with the statistical value of each point value in a 3 x 3 neighborhood of the point, and carrying out zero-taking processing on the vacant point, wherein the sound code signal intensity average value filtering formula is as follows:
(x, y) represents the position of a point in the two-dimensional sound code signal, g (x, y) is the statistical value of the point, f (x-k, y-l) is the value of the point (x-k, y-l), n is the size of the template, k and l take the values of-1, 0,1, and W is a two-dimensional sliding template of 3 x 3.
The present invention will be described in detail with reference to examples.
Examples
With reference to fig. 1, a method for detecting the type of an acoustic code of a digital interphone based on signal two-dimensional recombination fusion filtering specifically comprises the following steps:
step 1, firstly, performing signal two-dimensional recombination on a one-dimensional acoustic code signal, and dividing the one-dimensional acoustic code signal originally containing 189 bytes according to 27 bytes of each line.
Assume that there are N different samples: x ═ X1,x2,…,xN|xi∈RD,i∈[1,2,…,N]},Y={y1,y2,…yi…,yN1,2, …, N, where D is the data size of each sample of the one-dimensional vocoded signal, the size is 189 bytes, and N is the number of samples; y is a set of one-dimensional vocoded signal labels, YiIs the label of the ith one-dimensional vocoded signal.
Because the one-dimensional vocoded signals are easily affected by noise and have obvious intra-class differences, the characteristics of the input signals cannot be effectively extracted. To make more full use of the sample information, we choose to use a method that recombines the one-dimensional signals in two dimensions to recombine the samples. By dividing the original one-dimensional vocoded signal containing 189 bytes by 27 bytes per line, a two-dimensional signal with 7(189/27) lines is obtained, and the one-dimensional sample information is converted into two-dimensional sample information, where the converted sample can be expressed as:
X'={x11,x12,…,xNT|xij∈RD/T,i∈[1,2,…,N],j∈[1,2,…,T]where T is the number of rows of the two-dimensional signal into which each one-dimensional signal is divided. After the division is finished, two-dimensional sample data x of N x T rows and D/T columns are obtainedijIs the jth line of the ith two-dimensional sound code signal.
And 2, preprocessing the recombined two-dimensional signal by using a fusion filtering method.
And preprocessing the two-dimensional recombined sound code signal mainly by a sound code signal intensity mean value filtering method in the fusion filtering process. Firstly, a 3 x 3 two-dimensional sliding template is constructed, the value of one point in the two-dimensional sound code signal is replaced by the statistic value of each point value in a 3 x 3 neighborhood of the point, and the method can effectively inhibit noise. The mean value filtering formula of the sound code signal intensity in the fusion filtering process is as follows
g(x,y)=∑(f(x-k,y-l))/n,(k,l∈W)
(x, y) represents the position of a point in the two-dimensional sound code signal, g (x, y) is the statistical value of the point, f (x-k, y-l) is the value of the point (x-k, y-l), n is the size of the template, k and l take the values of-1, 0,1, and W is a two-dimensional sliding template of 3 x 3.
And 3, selecting the characteristics capable of effectively distinguishing different types of signals by using an attribute reduction method aiming at the two-dimensional sound code signal characteristics preprocessed by the fusion filtering method, and compressing a high-dimensional characteristic space to a low-dimensional characteristic space. And taking the feature data obtained by the attribute reduction method as input, using a Support Vector Machine (SVM) as a classifier, and processing the extracted features to obtain an SVM classification model.
In order to reduce the complexity of calculation and improve the accuracy of classification identification, the invention selects a method using attribute reduction. According to the characteristics of the digital interphone signals, a plurality of signal characteristic types are selected for analysis, wherein the signal characteristic types comprise signal bandwidth, mean value, variance, peak point, frequency point occupancy, normalized instantaneous amplitude absolute value mean square error, normalized instantaneous amplitude kurtosis and frequency spectrum signal deviation. And recording all the extracted features as feature sets, selecting the most effective features from the feature sets by applying an attribute reduction method, and compressing a high-dimensional feature space to a low-dimensional feature space to realize high-precision detection of different acoustic code signal types.
The invention uses SVM classifier to classify the extracted sound code signal characteristics. The kernel function selection, the penalty factor C and the kernel function parameter g in the SVM classifier directly influence the SVM induction performance. The kernel function mainly has the function of mapping the nonlinear sample data to a high-dimensional space, so that linear classification is realized in the high-dimensional space. In order to obtain the optimal classification result and avoid the influence of different kernel functions on the classification effect, the RBF kernel function is proved to obtain higher classification precision through test comparison, so the kernel function adopted in the method is the RBF radial basis function. The penalty factor C is used for representing the degree of importance on outliers, the kernel function parameter g represents the radius of a kernel function, the two parameters are mainly trained by a network cross verification method (Grid method for short), some possible C and g values are selected according to experience, and the parameter with the highest classification precision is selected by a k-cross method in a preferred mode. In this embodiment, the value of the parameter C is 100, and the value of the parameter g is 1 e-5.
And 4, carrying out two-dimensional signal recombination and fusion filtering processing on the digital interphone sound code signal sample to be detected.
And 5, selecting features capable of effectively distinguishing signals of different classes by using an attribute reduction method for the acoustic code signal sample preprocessed in the step 4, predicting the class corresponding to each feature of the acoustic code signal by using the SVM classification model obtained in the step 3, performing decision fusion on the classification result predicted by each feature of the same sample by using a voting method, improving the classification precision, and obtaining a final result corresponding to the sample.
The method uses an SVM classification model to predict the category corresponding to each feature after the attribute of the acoustic code signal to be detected is reduced, then carries out decision fusion through a voting method, and classifies the test sample into a class with the same decision obtained by most features.
Let Xt={x1,x2,…,xT|xj∈RD,j∈[1,2,…,T]Denotes the two-dimensional characteristics of a single test sample of the vocoded signal, Yt={y1,y2,…,yT′,Representing the classification result corresponding to each feature after attribute reduction of a single acoustic code signal test sample, wherein T' is the dimension after attribute reduction; the expression for decision fusion using voting is as follows:
wherein, count (p) indicates that the category p is in the sample label YtC is the set of possible classes of vocoded signals.
Claims (4)
1. The utility model provides a digit intercom sound code type detection method based on signal two-dimensional recombination fuses filtering which characterized in that includes following step:
step 1, performing two-dimensional signal recombination on a one-dimensional acoustic code signal to obtain a two-dimensional acoustic code signal; the method specifically comprises the following steps:
suppose there isNThe different samples:whereinDFor the amount of data of each sample of the one-dimensional vocoded signal,in order to be a sample of the sample,Nis the number of samples;
converting the one-dimensional sample information into two-dimensional sample information, wherein the converted sample is represented as:
whereinIs as followsSecond of two-dimensional sound code signalThe rows of the image data are, in turn,Tdividing the number of lines of the two-dimensional signal into each one-dimensional signal to obtain the number of lines of the two-dimensional signalThe rows of the image data are, in turn,two-dimensional sample data of a column;
step 2, preprocessing the recombined two-dimensional sound code signal by using a fusion filtering method;
step 3, using a support vector machine as a classifier, and performing model training on the two-dimensional acoustic code signals subjected to fusion filtering according to features obtained by attribute reduction to obtain an SVM classification model;
step 4, carrying out two-dimensional signal recombination and fusion filtering processing on the digital interphone sound code signal sample to be detected;
and 5, testing and detecting the characteristics of the two-dimensional sound code signals subjected to fusion filtering processing according to attribute reduction by using the SVM classification model obtained by training in the step 3, and performing decision-level fusion on the test detection result of each two-dimensional sound code signal by adopting a voting method to obtain a final sound code signal type detection result.
2. The method for detecting the type of the vocoder of the digital interphone based on the signal two-dimensional recombination fusion filtering as claimed in claim 1, wherein the step 1 comprisesDOne-dimensional vocoded signal of one byte by each lineByte divisionTA line whereinDIs composed ofTInteger multiples of.
3. The method for detecting the type of the digital interphone vocoded based on the signal two-dimensional recombination fusion filtering as claimed in claim 2,Dthe value of (a) is 189,Tis 7.
4. The digital interphone vocoded type detection method based on signal two-dimensional recombination fusion filtering of claim 1, wherein in the step 2, the fusion filtering is to preprocess the vocoded signal after two-dimensional recombination by a vocoded signal intensity mean filtering method;
firstly, constructing a 3 x 3 two-dimensional sliding template, replacing the value of one point in a two-dimensional sound code signal with the statistical value of each point value in a 3 x 3 neighborhood of the point, and carrying out zero-taking processing on the vacant point, wherein the sound code signal intensity average value filtering formula is as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810074257.XA CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810074257.XA CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108417221A CN108417221A (en) | 2018-08-17 |
CN108417221B true CN108417221B (en) | 2021-09-21 |
Family
ID=63126159
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810074257.XA Active CN108417221B (en) | 2018-01-25 | 2018-01-25 | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108417221B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1064562B1 (en) * | 1998-03-17 | 2009-01-14 | QUALCOMM Incorporated | System and method for determining the position of a wireless cdma transceiver |
US8213953B1 (en) * | 2004-05-11 | 2012-07-03 | Sprint Spectrum L.P. | Method and system for vocoder bypass based on loads in coverage areas of a wireless wide area network |
CN102763160A (en) * | 2010-02-18 | 2012-10-31 | 高通股份有限公司 | Microphone array subset selection for robust noise reduction |
CN206461605U (en) * | 2017-01-04 | 2017-09-01 | 泉州力同科技有限公司 | A kind of integrated talkback module and its system |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6418000B2 (en) * | 2015-02-19 | 2018-11-07 | アイコム株式会社 | Wireless communication system, wireless device, wireless communication method, packet generation method, and data reproduction method from packet |
DE102015005577A1 (en) * | 2015-05-04 | 2016-11-10 | Dräger Safety AG & Co. KGaA | Head protection device, communication unit, connection unit and system of head protection device, communication unit and connection unit |
-
2018
- 2018-01-25 CN CN201810074257.XA patent/CN108417221B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1064562B1 (en) * | 1998-03-17 | 2009-01-14 | QUALCOMM Incorporated | System and method for determining the position of a wireless cdma transceiver |
US8213953B1 (en) * | 2004-05-11 | 2012-07-03 | Sprint Spectrum L.P. | Method and system for vocoder bypass based on loads in coverage areas of a wireless wide area network |
CN102763160A (en) * | 2010-02-18 | 2012-10-31 | 高通股份有限公司 | Microphone array subset selection for robust noise reduction |
CN206461605U (en) * | 2017-01-04 | 2017-09-01 | 泉州力同科技有限公司 | A kind of integrated talkback module and its system |
Non-Patent Citations (5)
Title |
---|
GPU Parallel Implementation of Spatially Adaptive Hyperspectral Image Classification;Zebin Wu et al;《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》;20171031;全文 * |
SIP in 3G Wireless Networks: Service Models,Architecture, and Network Design;Carlos Urrutia-Valdés et al;《Bell Labs Technical Journal》;20041231;第9卷(第3期);全文 * |
基于AMBE声码器的数字对讲无线语音传输系统研究;韦建超;《中国优秀硕士学位论文全文数据库 信息科技辑》;20100215(第02期);全文 * |
基于数字对讲机的MELP声码器算法研究与实现;黄慧猛;《中国优秀硕士学位论文全文数据库 信息科技辑》;20120515(第05期);全文 * |
基于迭代加权L1 正则化的高光谱混合像元分解;吴泽彬等;《南京理工大学学报》;20110831;第35卷(第4期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN108417221A (en) | 2018-08-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111292764B (en) | Identification system and identification method | |
CN101546556B (en) | Classification system for identifying audio content | |
CN109242209B (en) | Railway emergency grading early warning method based on K-means clustering | |
CN109936857B (en) | Intelligent identification method for wireless perceptibility | |
CN113762377B (en) | Network traffic identification method, device, equipment and storage medium | |
JP6783443B2 (en) | Information processing equipment, information processing systems, information processing methods, programs, and recording media | |
CN113077444A (en) | CNN-based ultrasonic nondestructive detection image defect classification method | |
CN110751191A (en) | Image classification method and system | |
CN107274912B (en) | Method for identifying equipment source of mobile phone recording | |
CN113488063A (en) | Audio separation method based on mixed features and coding and decoding | |
CN114151293B (en) | Fault early warning method, system, equipment and storage medium of fan variable pitch system | |
CN116910752A (en) | Malicious code detection method based on big data | |
CN114022923A (en) | Intelligent collecting and editing system | |
CN116705059B (en) | Audio semi-supervised automatic clustering method, device, equipment and medium | |
CN108417221B (en) | Digital interphone sound code type detection method based on signal two-dimensional recombination fusion filtering | |
CN112817954A (en) | Missing value interpolation method based on multi-method ensemble learning | |
CN110310660B (en) | Speech resampling detection method based on spectrogram | |
WO2021088176A1 (en) | Binary multi-band power distribution-based low signal-to-noise ratio sound event detection method | |
CN106372655A (en) | Synthetic method for minority class samples in non-balanced IPTV data set | |
CN111368858A (en) | User satisfaction evaluation method and device | |
CN115062111A (en) | Fault duplicate report generation method and device and electronic equipment | |
CN114330720A (en) | Knowledge graph construction method and device for cloud computing and storage medium | |
CN104166855A (en) | Method of visual voice recognition by following-up the local deformations of a set of points of interest of the speaker's mouth | |
CN114662529A (en) | Binary tree filter Transformer model and bearing fault diagnosis method thereof | |
CN109935234B (en) | Method for identifying source equipment of sound recording |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |