33 results about "Multiscale entropy" patented technology

In a subject undergoing therapeutic intervention, efficacy of the therapeutic intervention is assessed based on a series of physiologic data associated with the subject. The series of physiologic data is analyzed to produce a measure of complexity. The complexity measure is then compared to a control. The efficacy of the therapeutic intervention is assessed based on the comparison of the complexity measure to the control. The control may be, for example, a complexity measure taken prior to initiation of the therapeutic intervention, a complexity measure taken from a different subject, or a predetermined threshold value. The measure of complexity is generated using, for example, a multiscale entropy measurement (MSE), a time asymmetry measurement, and / or an information-based similarity measurement. An increase in complexity indicates a positive effect of the therapeutic intervention, while a decrease in complexity indicates a negative effect of the therapeutic intervention. Stability of a non-biologic network, such as a computer network, communications network or transportation network can also be assessed.

The invention belongs to the field of medical technology, and relates to a method for analyzing heart rate variability data based on self-adaptive multi-scale entropy; the invention overcomes the shortcomings of MSE quantifying HRV complexity in a fixed scale, and quantifies HRV complexity comprehensively and accurately; The HRV is adaptively decomposed by state decomposition, and a series of multi-scale mean substitution data sets of the original data are obtained; secondly, each element in the mean substitution data set is coarse-grained and the corresponding adaptive scale is obtained; finally, the coarse-grained calculation is carried out. The averaged value replaces the SampEn value of each element in the data set, and the adaptive multi-scale entropy is obtained; the adaptive multi-scale entropy can comprehensively and accurately evaluate the HRV complexity.

The invention relates to a method for locating pipeline multi-point leaks based on improved VMD, comprising: collecting original pipeline leakage signals; performing overall local mean value decomposition on the original leakage signals to obtain several PF components; calculating the correlation coefficient of each PF component, Screen out the required PF component according to the correlation coefficient, and carry out signal reconstruction according to the screened out PF component, determine the k value of the variational mode decomposition; carry out the variational mode decomposition to the reconstructed signal, obtain several IMF components, Calculate the multi-scale entropy value of each IMF component, and screen the IMF component according to the multi-scale entropy value of each IMF component; perform signal reconstruction on the screened IMF component, and perform cross-correlation positioning calculation on each leakage signal after blind source separation Complete pipeline leak location. The invention can effectively extract pipeline leakage signals, eliminate the influence of low correlation components and noise in the original leakage signals, and make the final positioning result more accurate.

The invention discloses an industrial system fault detection method based on Euclidean distance multiscale fuzzy sample entropy. According to the method provided by the invention, the complexity of atime sequence can be described from a plurality of time scales; meanwhile, compared with an existing multiscale entropy method and an existing composite multiscale entropy (FME) method, the method hasthe advantages that the stability and the accuracy of the calculation of multi-scale fuzzy sample entropy (FME) are remarkably improved. The method can be used for judging and detecting the fault type of an industrial system and analyzing the complexity of the time sequence.

The invention discloses an image quality evaluation method based on multi-scale gradient difference entropies, and the method comprises the following steps: 1, converting an original image and a distorted image into gray-scale images; 2, calculating gradient difference entropies in different scale spaces based on the results obtained at step 1, and enabling the gradient difference entropies in different scale spaces to be averaged in a weighted manner; 3, calculating the ration of mean values of the gray-scale images corresponding to the original image and the distorted image; 4, calculating the ratio of variances of the gray-scale images corresponding to the original image and the distorted image; 5, carrying out comprehensive evaluation based on the results obtained at step 2, 3, and 4. The method introduces a basic concept of scale space, is based on a forming process of a target on a retina of a person at different distances, and is better in consistency with subjective evaluation results through the extraction of contour characteristics in the scale space and the reflecting of the difference between contour characteristics of the original image and the distorted image in the scale space through the entropies.

The invention discloses a mental workload online detection method based on forehead electroencephalogram signals. The method comprises the following steps that the forehead electroencephalogram signals are collected with a silver / silver chloride electrode as a sensor; an electroencephalogram amplifier is used for carrying out amplification and filtering on the forehead electroencephalogram signals, then data preprocessing is carried out, and the processed forehead electroencephalogram signals are acquired; a stimulation task is compiled as n-back; wavelet multiscale entropy features are extracted from the processed forehead electroencephalogram signals; model recognition is carried out on the wavelet multiscale entropy features through a support vector machine, and an obtained result is the mental workload level and recognition accuracy of data. The experiment process of the method is carried out on the forehead, the need for washing the head before and after the process is avoided, the operation is facilitated, and meanwhile the influence on signal collection by the hair and the scalp is avoided. The method can effectively improve the accuracy and convenience of a mental workload detection system and obtain considerable social benefits and economic benefits.

The invention discloses a biological tissue terahertz imaging method, system and equipment based on multiscale entropy. According to the method, the information of whole terahertz time-domain spectroscopy signals is utilized, and the entropy value of each point on the detected biological tissue sample on different time scales is obtained; the entropy value on a certain scale is selected, or the entropy values on multiple scales are added together, so that a terahertz image of the detected biological tissue sample is reconstructed. By means of the method, the limitations that a traditional biological tissue terahertz imaging method only uses the spectroscopy information of a certain time or frequency point are overcome, and imaging contrast among different samples is improved.

The invention provides a pipeline leakage signal identification method based on improved ELMD multi-scale entropy. The pipeline leakage signal identification method includes the steps that firstly, acquired experimental data are subjected to noise pre-processing to eliminate low-correlation components in signals; then pre-processed signals are subjected to ELMD processing to obtain PF components;an endpoint effect problem reserved by ELMD decomposition is weakened through a peak value waveform matching method; the multi-scale entropy of the PF components is calculated correspondingly, and multi-scale entropy values of leakage signals are arrayed and compared to eliminate background noise; a feature vector is constructed by selecting a principal PF component according to the multi-scale entropy values; the feature vector is used as an input vector of a BP neural network to train the network; and a to-be-tested sample is input into a trained BP neural network to obtain a pipeline leakage identification result. The pipeline leakage signal identification method based on the improved ELMD multi-scale entropy can adapt to various conditions of pipelines and has good testing accuracy.

The invention relates to a motor bearing fault diagnosis method based on feature optimization and GWOA-XGBoost, comprising the following steps: using optimized VMD to decompose the vibration signal of the motor bearing to obtain its IMF component, and calculating the IMF component based on the multi-scale entropy theory MSE value, reconstruct the characteristics of the vibration signal through the MSE value; use the GA algorithm to improve the WOA algorithm to obtain the GWOA model, optimize the XGBoost model through the GWOA model; input the characteristic reconstructed vibration signal into the optimized XGBoost model , to obtain the motor bearing fault diagnosis results. The motor bearing fault diagnosis method of the invention can improve the accuracy of the motor bearing fault diagnosis.

The invention discloses a sound signal noise reduction method and system of foreign matter in an electric energy meter on the basis of short-time energy, multi-scale entropy and EMD. The method includes the steps that noise data generated by shaking of the foreign matter in the electric energy meter is collected; foreign matter sound signal data is extracted through short-time energy positioning;the extracted foreign matter sound signal data is subjected to empirical mode decomposition (EMD), and the multi-scale entropy of each stage of intrinsic mode function (IMF) component obtained throughEMD is calculated; the multi-scale entropy of the IMF components is subjected to noise reduction and filtering to obtain sound signals, obtained after noise reduction and filtering, of the foreign matter in the electric energy meter. The method has the advantages that short-time energy positioning can be used for data extraction to increase the processing speed; the sound signals of the foreign matter in the electric energy meter are processed through multi-scale entropy and empirical mode decomposition, which is a self-adaptive process, and the defect that traditional spectrum subtraction noise reduction estimation is not accurate is overcome; meanwhile, introduction of new noise is avoided to the maximum degree, and the foreign matter in the electric energy meter can be accurately detected.