47 results about "Spectral centroid" patented technology

A mobile station has a processor and a user input mechanism such as a button that is operable to cause the processor to extract at least one feature from a digital media sample. The feature, such as a spectral centroid, is descriptive of an identity of a content of the media sample. Preferably, the same user input that causes the feature extraction also causes a transmitter to establish a link and to send a message to a network address, the message having a plurality of extracted features from which the digital media sample may not be reconstructed. Where a reply message identifies a media file that matches the media sample, another user input at the same or a different button causes an authorization message to be sent so that a copy of the media file, identified in the reply message and having features that exactly match the plurality of extracted and transmitted features, is downloaded to the mobile station.

The invention discloses a method for classifying digital audio automatically. The method particularly comprises the following steps: pretreating an audio signal: pre-emphasis treatment, framing treatment and windowing treatment; extracting an audio characteristics comprising frequency spectrum mass center, spectrum spreading, frequency spectrum flatness, frequency spectrum transition parameters, short time energy, fundamental frequency and Mel frequency cepstrum coefficient (MFCC) and MFCC first order difference; indexing and classifying according to the audio characteristic, and finding out the similar audio. According to the invention, the problem of error caused by indexing audio by adopting a single audio characteristic in the prior art is solved; moreover, the method is simple in computation process, is easy to apply practically, and is high in indexing efficiency.

A method for sensing touch inputs to a digital equipment is provided, comprising the steps of sensing a sound / vibration signal generated by a touch, digitally processing the sensed sound / vibration signal, and determining the type of touch means that has generated the touch and the intensity of the touch based on the properties of the processed sound / vibration signal, wherein the properties include at least one of the following properties of the sound / vibration signal in time domain: maximum amplitude, average amplitude, average frequency, mean, standard deviation, standard deviation normalized by overall amplitude, variance, skewness, kurtosis, sum, absolute sum, root mean square (RMS), crest factor, dispersion, entropy, power sum, center of mass, coefficients of variation, cross correlation, zero-crossings, seasonality, DC bias, or the above properties computed for the first, second, third or higher order of derivatives of the sound / vibration signal; and the following properties of the sound / vibration signal in frequency domain: spectral centroid, spectral density, spherical harmonics, total average spectral energy, band energy ratios for every octave, log spectral band ratios, linear prediction-based cepstral coefficients (LPCCs), perceptual linear prediction (PLP) cepstral coefficients, mel-frequency cepstral coefficients, frequency topology, or the above properties computed for the first, second, third or higher order of derivatives of a frequency domain representation of the sound / vibration signal. There is also provided a device for sensing touch inputs.

The invention belongs to the technical field of medical ultrasonography, and particularly relates to a cancellous bone diagnosis system based on ultrasound backscattering signal parameters. The system comprises an ultrasound backscattering signal acquisition module, a pretreatment module, a backscattering parameter calculation module and a cancellous bone state evaluation module. In the invention, an ultrasound transceiver probe is used to acquire ultrasound backscattering signals from the calcaneus of a human body; effective cancellous bone ultrasound backscattering signals are extracted; the calculation module is used to calculate four parameters, i.e. backscattering coefficient, apparent integration backscattering coefficient, spectral centroid shift and mean trabecular bone spacing; and finally, the health state of the cancellous bone is analyzed by comparing the four parameters with standard values stored in an internal standard database in the system. Compared with the traditional ultrasound transmission diagnosis system based on broadband ultrasound attenuation and ultrasound propagation velocity, the diagnosis system provided by the invention can obtain the complete information of the cancellous bone structure, thereby better detecting the health state of the cancellous bone of a human body.

A method for optical detection of characteristic quantities of the wavelength-dependent behavior of an illuminated specimen, such as the emission behavior and / or absorption behavior, preferably the fluorescence and / or enzyme fluorescence comprises determining at least one spectral centroid and / or a maximum of the emission radiation and / or of the absorbed radiation. An arrangement for performing the method is disclosed.

The invention discloses an audio file retrieval method and an audio file retrieval device and belongs to the field of audio processing. The method comprises the steps of collecting a to-be-retrieved audio fragment in real time; acquiring a difference value of frequency spectrum centroids of corresponding sub-bands of every two audio frames of adjacent time sequences in a plurality of audio framesof the audio fragment to obtain a plurality of difference values; obtaining an audio fingerprint corresponding to the two audio frames based on the plurality of difference values and a preset difference value threshold, wherein each bit of the audio fingerprint is determined by whether the difference value corresponding to the bit meets the preset condition of the difference value; and retrievingin an audio file library based on the audio fingerprint of multiple audio frames to obtain a target audio file matched with the audio fragment. According to the audio file retrieval method and the audio file retrieval device, the accuracy for retrieval of the audio file is improved.

The invention provides a high-frequency oscillation rhythm detection method based on an optimal fuzzy clustering Gaussian mixture model. A method based on clustering analysis is adopted to detect thehigh-frequency oscillation rhythm, fuzzy entropy, short-time energy, a power ratio and a frequency spectrum center of mass are selected as the characteristics of an epilepsy electroencephalogram, thecharacteristics are formed into a characteristic vector as the input of a clustering algorithm, an expectation-maximization Gaussian mixture model clustering algorithm is adopted to classify the characteristic vector, and a fuzzy c-mean clustering algorithm is adopted to obtain the initialization parameter of the expectation-maximization Gaussian mixture model clustering algorithm; and a median and a quartile range are selected to analyze each category of statistics characteristics, and the high-frequency oscillation rhythm is detected. The method has the beneficial effects that the detectionspeed of the high-frequency oscillation rhythm of the epilepsy electroencephalogram is improved, and doctors are assisted in carrying out epilepsy diagnosis and epileptogenic focus excision.

The invention provides a respiration and heartbeat signal separation method and device based on blind source separation. The method comprises the steps that 1, time frequency domain analysis is conducted on all biological radar echo signals, and time-frequency distribution signals of biological radar echoes corresponding to all targets are obtained; 2, on the basis of the time-frequency distribution signals, frequency spectrum centroid curves of all the biological radar echoes are obtained; 3, all the frequency spectrum centroid curves serve as observed signals, blind source separation is conducted, and micro change signals corresponding to all the targets are obtained; 4, filtering is conducted on the micro change signals, and respiratory time-varying signals and heartbeat time-varying signals which correspond to all the targets are obtained. According to the respiration and heartbeat signal separation method and device based on blind source separation, the biological radar echo signals of two target human bodies can be processed, and the respiration and heartbeat signals for each target are separated out.

The invention discloses a typical abnormal sound alarm method for an elevator based on a depth single classifier. The method comprises the following steps of step 1, respectively extracting RMS features of five typical abnormal sound signals; step 2, respectively extracting spectral centroid features of the five typical abnormal sound signals; step 3, respectively extracting MFCC features of the five typical abnormal sound signals; step 4, performing feature extraction on input training samples; step 5, computing an output weight beta of classified learning; step 6, computing a single classifier threshold value theta; and step 7, inputting test data for testing. According to the method, the reaction of a trapped person in the elevator can be rapidly and accurately detected and identified,and through linkage with an automatic alarm system, rapid alarm is facilitated, and the personnel can be notified in time. A guarantee is provided for safe and efficient operation of the elevator, thetimely treatment of the danger and the reduction of the loss of life and property.

The invention discloses an intrusion signal diagnostic system and an intrusion signal diagnostic method. The intrusion signal diagnosis system and the intrusion signal diagnosis method are characterized in that an original signal is collected, and an EEMD is used for decomposing the original signal into an IMF component, then performing de-noising reconstruction, and extracting a spectral centroidof the reconstructed signal through a spectral centroid method, and the spectral centroid is used as a characteristic quantity to be imported into the LS-SVM to carry out regression prediction, and therefore the intrusion signal is diagnosed.

The embodiment of the invention provides a creation method of a body weight acoustic measurement model, a body weight measuring method, a creation device of the body weight acoustic measurement model,and a body weight acoustic measuring device. The creation method of the body weight acoustic measurement model comprises the following steps of calculating acoustic feature parameters of animal audiodata, wherein the acoustic feature parameters include the sound pressure, the peak value frequency, the frequency spectrum centriod and the power spectrum; obtaining the real body weight of an animalowning the audio data; creating the weight acoustic measurement model; and training the weight acoustic measurement model by using the acoustic feature parameters of the animal audio data and the real body weight of the animal owning the audio data to obtain a body weight estimated parameter value in the body weight acoustic measurement model. The creation method of the body weight acoustic measurement model, the body weight measuring method, the creation device of the body weight acoustic measurement model, and the body weight acoustic measuring device provided by the embodiment of the invention have the advantages that the body weight of the animal can be estimated through sound; firm foundation is provided for real-time fast low-cost and non-contact measurement of the body weight of the animal; and wide application prospects are realized.

An ophthalmological measuring system for determining distances and / or for tomographic imaging of ocular structures, based on an OCT method. The measuring system includes a light source with a spectral centroid (λ), an interferometric measuring device, a scanner system, which in addition to the lateral deflection of the sample beam also has axial modulations with a frequency (f) in the sample arm, and a control and evaluation unit. The scanner performs a lateral, two-dimensional deflection of the sample beam with the aid of one or even two separate mirror elements and can in particular have axial modulation amplitudes zM>>λ / 2. The system can also be used for scanner systems in other fields that use an OCT method, in particular a swept-source OCT method.

Increasing the level of the consonant segments relative to the nearby vowel segments, known as consonant-vowel ratio (CVR) modification, is reported to be effective in improving speech intelligibility by listeners in noisy backgrounds and by hearing-impaired listeners. A method along with a system for real-time CVR modification using the rate of change of spectral centroid for detection of spectral transitions is disclosed. A preferred embodiment of the invention using a 16-bit fixed point processor with on-chip FFT hardware is also presented for real-time signal processing. It can be integrated with other FFT-based signal processing in communication devices, hearing aids, and other systems for improving speech perception under adverse listening conditions.

The invention provides a target micro-Doppler component extraction method based on centroid calculation. The method includes the steps of the collection of target motion I and Q echo signals, data segmentation, orthogonalization processing, complex processing, FFT transformation, threshold value setting and noise floor removal, the removal of out-of-band noise by estimating a target frequency bandrange, frequency spectrum centroid solution, the frequency mixing with an original complex signal after the centroid is obtained, namely, the removal of interference of Doppler quantity generated bytarget translation to obtain characteristics of a target micro-motion. The method has the advantages of simple calculation, strong operability, high precision and high real-time performance and effectively overcomes the defects of low real-time performance, complex device and high priori information requirement of an existing algorithm.

According to one aspect of the invention, there is provided a method for calculating information on urination. The method includes the steps of acquiring a sound signal related to urination of a user, calculating a urination parameter related to an effective value (or urination parameters related to an effective value and a spectral centroid) from the acquired sound signal, and estimating a urine flow rate of the user with reference to the calculated urination parameter.

The invention provides a Gabor transformation based method and device for extracting and analyzing breathing and heartbeat signals. The method comprises following steps: S1, radar echo signals are subjected to Gabor transformation, and time-frequency power distribution signals of the transformed signals are calculated; S2, the time-frequency power distribution signals are subjected to frequency spectrum weighted average operation, and a frequency spectrum centroid changing curve is obtained; S3, the frequency spectrum centroid changing curve is subjected to filtering operation, and breathing signals and heartbeat signals are obtained; S4, the breathing signals and the heartbeat signals are subjected to the Gabor transformation, and breathing frequency changing with time and heartbeat frequency changing with the time are obtained. With the adoption of the method and the device, the breathing frequency and heartbeat frequency changing with time, the signals and the frequency change of the signals can be effectively extracted.

The invention provides a high-frequency oscillation rhythm detection method intelligent algorithm optimized fuzzy clustering to detect high-frequency oscillation rhythms on the basis of the fuzzy clustering method. The method comprises basic steps as follows: selecting a mean singular value (MSV), a wire length fl, a power ratio and a frequency spectrum centroid as features of an epilepsy electroencephalogram (EEG), and constructing the features into a feature vector to be input as a clustering algorithm; optimizing a fuzzy clustering algorithm by a simulated annealing genetic algorithm in anintelligent algorithm to obtain an optimized parameter vc; obtaining an optimized result according to the optimized parameter vc; and selecting medians and interquartile ranges to analyze statisticalfeatures of each class to detect the high-frequency oscillation rhythms. The method has the beneficial effects as follows: the detection precision of high-frequency oscillation rhythms of epilepsy EEGis improved and doctors are helped to perform epilepsy diagnosis and epileptogenic focus excision.

The invention provides a wind noise detection method, device and system. The method comprises that each frame of audio data is converted into frequency-domain data, and whether the audio data includeswind noise is determined according to the spectral centroid of the frequency-domain data. In one hand, wind noise detection can be realized via one path of audio data instead of collecting two pathsof audio data for comparison, and thus, operation is convenient; and in the other hand, compared with a scheme which needs two collectors to collecting two paths of audio data, only one collector is needed to collect one path of audio data, and the equipment cost is reduced.

The invention provides a synchronous information extraction method for a video leakage signal. The method comprises the following steps: the power spectrum of the video leakage signal is acquired and divided into multiple sub-bands; the spectral centroid of each sub-band is acquired, and an interval between the spectral centroids of every two sub-bands is acquired; line frequency or field frequency of the signal is acquired according to the uniform degree of the spaced distribution of the spectral centroids. The invention further provides a synchronous information extraction system for the video leakage signal. The system comprises a segmentation module, a calculation module and an extraction module. The line frequency or field frequency of the video signal can be extracted accurately and automatically in the presence of noise and interference signals.

The invention discloses an ultrasonic positioning method for a partial discharge fault point of a traction transformer. Ultrasonic signals are obtained and denoised through four ultrasonic sensors arranged on the surface of a transformer; the collected ultrasonic waveform is analyzed to obtain waveform characteristic parameters, wherein the waveform characteristic parameters comprise a peak-rise time ratio, kurtosis, peak frequency and a frequency spectrum centroid; a path judgment coefficient of the ultrasonic signal is obtained, wherein the judgment coefficient can identify a propagation path of the ultrasonic signal; and finally, a time difference equation is established and calculated to obtain a local discharge source coordinate. The method effectively improves the partial discharge positioning precision and positioning efficiency of the transformer.

The invention belongs to the technical field of medical instruments, and particularly relates to a back-scattering ultrasonic bone quality diagnosis system based on an Android platform. The system is composed of a hardware layer, a driving layer, an Android system layer and an application layer; the architecture combining ARM, an FPGA and an analog circuit is adopted in hardware of the bottom layer, and an Android system runs on an ARM processor and controls the hardware of the bottom layer through the driving layer; the application layer runs on the Android system to achieve all processes and functions of back-scattering ultrasonic bone quality diagnosis; the application layer comprises an ultrasonic measurement module, a parameter setting module, a man-machine interaction module, an algorithm processing module and a database access module, wherein the algorithm processing module obtains parameters such as the back-scattering coefficient (BSC), the apparent integral back-scattering coefficient (AIB) and back-scattering spectral centroid skewing (SCS) through calculation of a signal processing algorithm and evaluates the bone quality by synthesizing the parameters; multi-task scheduling processing is achieved through the Android operation system, and therefore the reliability and the real-time response performance of the system are guaranteed.

The invention discloses a compressor blade crack fault detection method, which comprises the following steps of: collecting acoustic emission signals of two channels of an air outlet of a compressor, and dividing the acoustic emission signals into training samples and test samples; extracting an acoustic emission feature, a time domain feature, a frequency domain feature and a spectrum centroid energy migration feature of the training sample; performing feature selection by using a mixed feature selection method, and determining an optimal feature subset; establishing a two-channel training sample and a test sample feature subset; and combining the two channel training sample feature subsets, training by using a long short-term memory neural network, and finally carrying out blade crack fault classification and detection on a test sample by using the trained long short-term memory neural network to realize crack fault detection of the compressor blade. The method is simple and easy to implement, compared with other existing fault feature and crack detection technologies, the spectral centroid energy migration feature capable of effectively reflecting the blade fault feature can be established, and mixed feature selection and compressor blade crack fault detection are achieved.

The invention provides a noise reduction method and device of an earphone, electronic equipment and a storage medium, and relates to the technical field of audio processing, in particular to the technical field of voice processing, voice recognition and the like. According to the specific implementation scheme, the correlation between sound signals collected by two microphones in the earphone is obtained, and the two microphones comprise a feed-forward microphone and a call microphone; acquiring first energy of a sound signal acquired by a feed-forward microphone; determining a spectrum centroid of the sound spectrum according to the sound signal; performing wind noise detection on the earphone according to the correlation, the first energy and the spectral centroid; and performing active noise reduction on the earphone in response to detection of existence of the wind noise. Therefore, the influence of the correlation, the energy and the spectral centroid on noise detection is comprehensively considered, the detection of the earphone on the low-frequency noise can be improved by introducing the spectral centroid, and the situation that the active noise reduction effect is influenced due to the fact that the correlation of the low-frequency signal is low and the low-frequency noise is misjudged is prevented.