CN110400360B - Sound wave transit time detection method based on full convolution neural network - Google Patents
Sound wave transit time detection method based on full convolution neural network Download PDFInfo
- Publication number
- CN110400360B CN110400360B CN201910677510.5A CN201910677510A CN110400360B CN 110400360 B CN110400360 B CN 110400360B CN 201910677510 A CN201910677510 A CN 201910677510A CN 110400360 B CN110400360 B CN 110400360B
- Authority
- CN
- China
- Prior art keywords
- signal
- neural network
- time
- convolution neural
- full convolution
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
- G06T11/005—Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
Abstract
The invention discloses a sound wave transit time detection method based on a full convolution neural network. Compared with the traditional transit time detection method which only can combine a plurality of characteristics of the input signal, the sound wave transit time detection method based on the full convolution neural network can automatically extract a plurality of characteristics of different levels, not only can utilize the local detail information of the input signal, but also can combine the overall trend of the input signal for analysis, thereby obtaining more accurate results. In addition, the invention adopts the simulation data which is easy to mark for training, does not need to manually mark huge data sets one by one, can save the detection time and reduce the detection cost.
Description
Technical Field
The invention relates to the technical field of ultrasonic CT sound velocity imaging in biomedical ultrasonography, in particular to a sound wave transit time detection method based on a full convolution neural network.
Background
Ultrasound CT imaging can provide three-dimensional images. The ultrasonic CT sound velocity imaging is one of ultrasonic CT imaging, and based on the characteristic that sound velocities of ultrasonic waves in different tissues are different, normal glands and cancerous tissues can be distinguished by reconstructing sound velocity distribution in mammary glands, and even different types of tumors can be distinguished. A critical step in the reconstruction is the acquisition of the transit time of the ultrasound signal, i.e. the time from transmission to reception of the ultrasound.
Currently, there are two main methods for acquiring the transit time, which are most commonly used in the field of ultrasound CT. One is an aic (initial information criterion) method, which infers a time window with a possible initial point through the distance between fixed sensors, then traverses each sampling point in the time window, divides the time window into two parts by using the sampling point, calculates the sum of entropies of two time periods, and takes the corresponding division point as the initial point of the signal when the sum of entropies is minimum. The method has weak anti-noise capability, and noise can generate great influence on the detection result. The other method is a cross-correlation (CC) method in which a signal obtained by passing ultrasonic waves through a pure water medium is cross-correlated with a signal obtained by passing ultrasonic waves through a target medium, and a cross-correlation function is used to determine a transit time, and in particular, a difference in transit time between the two signals is found by a maximum value of the cross-correlation function. The method mainly utilizes the waveform correlation of signals to judge, so that the method has strong anti-noise capability. However, once passing through a relatively complex medium, the signal shape changes, causing large fluctuations in the value of the correlation function, resulting in a large decrease in accuracy.
Disclosure of Invention
In view of this, the invention provides a full convolution neural network-based acoustic wave transit time detection method, which is used for solving the problems of poor noise robustness and low acoustic wave transit time detection accuracy when a medium with a complex structure is penetrated by the existing detection method.
Therefore, the invention provides a method for detecting the transit time of sound waves based on a full convolution neural network, which comprises the following steps:
s1: carrying out time window interception on a signal to be detected, keeping the length of interception consistent, recording the corresponding time of a time window starting point in the signal to be detected, and carrying out absolute value taking and normalization processing on the signal to be detected;
s2: inputting the processed signal to be detected into a trained full convolution neural network model to obtain an output function;
s3: and searching the time corresponding to the maximum value of the output function in the signal to be detected, wherein the sum of the time corresponding to the maximum value of the output function in the signal to be detected and the time corresponding to the time window starting point recorded in the time window intercepting process of the signal to be detected in the signal to be detected is the detected transition time.
In a possible implementation manner, in the method for detecting a transit time of a sound wave provided by the present invention, the training process of the full convolution neural network model includes the following steps:
s21: acquiring simulation data by utilizing a sound velocity region divided by the nuclear magnetic resonance image;
s22: marking the obtained simulation data to obtain an ultrasonic signal first arrival time, wherein the shape of a marked label adopts a one-dimensional Gaussian function, and the ultrasonic signal first arrival time is taken as the central position of the Gaussian function;
s23: calculating the range of the first arrival time of the ultrasonic signals according to the distance between the ultrasonic transducers, carrying out time window interception on the simulation data and the label according to the calculated range of the first arrival time of the ultrasonic signals, keeping the length of interception consistent, recording the corresponding time of a time window starting point in the simulation data, adding white noise in the simulation data, carrying out absolute value taking and normalization processing on the simulation data added with the white noise, and finishing the manufacture of a training set;
s24: establishing a full convolution neural network model, inputting the manufactured training set into the full convolution neural network model, and training the full convolution neural network model;
s25: and continuously adjusting the weight of the full convolution neural network model by utilizing back propagation, and updating the structural parameters of the full convolution neural network model to obtain the full convolution neural network model with the optimal global parameter matrix.
In a possible implementation manner, in the method for detecting a transit time of a sound wave provided by the present invention, in step S21, the method for obtaining simulation data by using a sound velocity region divided by a nuclear magnetic resonance image specifically includes the following steps:
s211: obtaining a plurality of nuclear magnetic resonance images, carrying out image segmentation on the nuclear magnetic resonance images, and dividing a sound velocity region according to a segmentation result;
s212: and solving a wave equation of the ultrasonic wave passing through the sound velocity region by using a finite element method to obtain simulation data.
In a possible implementation manner, in the method for detecting a transit time of a sound wave provided by the present invention, in step S22, the obtained simulation data is labeled to obtain an initial arrival time of an ultrasonic signal, a shape of a label labeled with a one-dimensional gaussian function is used, and the initial arrival time of the ultrasonic signal is taken as a center position of the gaussian function, which specifically includes the following steps:
s221: and measuring the first arrival time of the ultrasonic signal by using the noiseless characteristic of the simulation signal and adopting an AIC (advanced information communication) method, wherein the shape of the labeled label adopts a one-dimensional Gaussian function, and the first arrival time of the ultrasonic signal is taken as the central position of the Gaussian function.
According to the method for detecting the acoustic wave transit time, provided by the invention, the full convolution neural network is applied to the detection of the transit time in the medical ultrasonic CT, so that higher transit time detection precision and better noise robustness can be obtained. Compared with the traditional transit time detection method which only can combine a plurality of characteristics of the input signal, the sound wave transit time detection method based on the full convolution neural network can automatically extract a plurality of characteristics of different levels, not only can utilize the local detail information of the input signal, but also can combine the overall trend of the input signal for analysis, thereby obtaining more accurate results. In addition, the invention adopts the simulation data which is easy to mark for training, does not need to manually mark huge data sets one by one, can save the detection time and reduce the detection cost.
Drawings
Fig. 1 is a flowchart of a method for detecting a transit time of a sound wave based on a full convolution neural network according to an embodiment of the present invention;
fig. 2 is a flowchart of a training process of a full convolution neural network model in a full convolution neural network-based acoustic transit time detection method according to an embodiment of the present invention;
FIG. 3 is a region of sound velocity divided by a nuclear magnetic resonance image;
FIG. 4 is a signal shape of the obtained simulation data;
FIG. 5 is a signal shape of simulation data after time window truncation;
FIG. 6 is a signal shape of simulation data after white noise is added;
FIG. 7 is a signal shape of the simulation data after taking the absolute value and then normalizing;
FIG. 8(a) is the shape of the tag after the time window has been cut;
FIG. 8(b) is the signal shape after time window truncation;
FIG. 9 is a schematic structural diagram of the established full convolution neural network model;
fig. 10 is a second flowchart of a training process of a full convolution neural network model in a full convolution neural network-based acoustic transit time detection method according to an embodiment of the present invention;
fig. 11 is a third flowchart of a training process of a full convolution neural network model in the sound wave transit time detection method based on the full convolution neural network according to the embodiment of the present invention;
fig. 12 is a detection result of single piece of simulation data obtained by using the full convolution neural network-based acoustic wave transit time detection method provided by the embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only illustrative and are not intended to limit the present application.
The method for detecting the acoustic wave transit time based on the full convolution neural network, as shown in fig. 1, includes the following steps:
s1: carrying out time window interception on the signal to be detected, keeping the length of interception consistent, recording the corresponding time of the starting point of the time window in the signal to be detected, and carrying out absolute value taking and normalization processing on the signal to be detected;
s2: inputting the processed signal to be detected into a trained full convolution neural network model to obtain an output function;
s3: and finding the time corresponding to the maximum value of the output function in the signal to be detected, wherein the sum of the time corresponding to the maximum value of the output function in the signal to be detected and the time corresponding to the time window starting point recorded in the time window intercepting process of the signal to be detected in the signal to be detected is the detected transit time.
According to the method for detecting the transit time of the sound wave, provided by the embodiment of the invention, the full convolution neural network is applied to the detection of the transit time in the medical ultrasonic CT, so that higher transit time detection precision and better noise robustness can be obtained. Compared with the traditional transit time detection method which only can combine a plurality of characteristics of the input signal, the sound wave transit time detection method based on the full convolution neural network can automatically extract a plurality of characteristics of different levels, not only can utilize the local detail information of the input signal, but also can combine the overall trend of the input signal for analysis, thereby obtaining more accurate results. In addition, the invention adopts the simulation data which is easy to mark for training, does not need to manually mark huge data sets one by one, can save the detection time and reduce the detection cost.
It should be noted that, in the above method for detecting acoustic wave transit time provided in the embodiment of the present invention, if the transit time difference needs to be measured, in step S2, the processed signal to be detected is input into the trained fully-convolutional neural network model to obtain an output function, and then the time corresponding to the maximum value of the output function in the signal to be detected is found, the time corresponding to the maximum value of the output function in the signal to be detected does not need to be added to the time corresponding to the start point of the time window in the signal to be detected, but the time corresponding to the maximum value of the output function in the input signal obtained from the input signal using pure water as a medium is subtracted, which is the transit time difference.
In specific implementation, in the method for detecting a transit time of a sound wave provided by the embodiment of the present invention, as shown in fig. 2, a training process of a full convolution neural network model may include the following steps:
s21: acquiring simulation data by utilizing a sound velocity region divided by the nuclear magnetic resonance image;
specifically, the division result is shown in fig. 3, the signal shape of the obtained simulation data is shown in fig. 4, and the received sampling frequency is set to be 10 MHZ;
s22: marking the obtained simulation data to obtain the first arrival time of the ultrasonic signal, wherein the shape of the marked label adopts a one-dimensional Gaussian function, and the first arrival time of the ultrasonic signal is taken as the central position of the Gaussian function;
s23: calculating the range of the first arrival time of the ultrasonic signals according to the distance between the ultrasonic transducers, carrying out time window interception on the simulation data and the label according to the range of the first arrival time of the ultrasonic signals, keeping the length of interception consistent, recording the corresponding time of the starting point of the time window in the simulation data, adding white noise in the simulation data, carrying out absolute value taking and normalization processing on the simulation data added with the white noise, and finishing the manufacture of a training set;
specifically, the signal shape of the simulation data after the time window interception is shown in fig. 5, the signal shape of the simulation data after white noise is added is shown in fig. 6, the signal shape of the simulation data after absolute value extraction and normalization processing is shown in fig. 7, the shape of the label after the time window interception is shown in fig. 8(a), the signal shape after the time window interception is shown in fig. 8(b), and fig. 8(a) and 8(b) show the corresponding relationship between the first arrival time of the ultrasonic signal in the simulation data and the center position of the label shape;
s24: establishing a full convolution neural network model, inputting the manufactured training set into the full convolution neural network model, and training the full convolution neural network model;
specifically, the structure of the established full convolution neural network model is shown in fig. 9;
s25: and continuously adjusting the weight of the full convolution neural network model by utilizing back propagation, updating the structural parameters of the full convolution neural network model, and obtaining the full convolution neural network model with the optimal global parameter matrix.
In specific implementation, in the method for detecting a transit time of a sound wave provided by the embodiment of the present invention, in step S21 in the training process of the full convolution neural network model, the sound velocity region divided by the nuclear magnetic resonance image is used to obtain simulation data, as shown in fig. 10, the method may specifically include the following steps:
s211: obtaining a plurality of nuclear magnetic resonance images, carrying out image segmentation on the nuclear magnetic resonance images, and dividing a sound velocity region according to a segmentation result;
s212: and solving a wave equation of the ultrasonic wave after the ultrasonic wave passes through the sound velocity region by using a finite element method to obtain simulation data.
In specific implementation, in the method for detecting acoustic wave transit time provided in the embodiment of the present invention, in step S22 in the training process of the full convolution neural network model, the obtained simulation data is labeled to obtain the first arrival time of the ultrasonic signal, the labeled label shape adopts a one-dimensional gaussian function, and the first arrival time of the ultrasonic signal is taken as the central position of the gaussian function, as shown in fig. 11, the method specifically includes the following steps:
s221: the characteristic that the simulation signal is noiseless is utilized, the ultrasonic signal first arrival time is measured by adopting an AIC method, the shape of the labeled label adopts a one-dimensional Gaussian function, and the ultrasonic signal first arrival time is taken as the central position of the Gaussian function. Specifically, according to the characteristic that simulation data almost has no noise, the initial arrival time of the ultrasonic signal can be detected by using a traditional AIC method, the initial arrival time is used as the real time of the initial arrival of the ultrasonic signal to manufacture a label of a training set, the shape of the label is a one-dimensional Gaussian function, the initial arrival time of the ultrasonic signal is used as the center position of the Gaussian function, and the sequence length of the label is the same as the sequence length of the input signal.
Fig. 12 is a detection result of a piece of simulation data, and it can be seen that the detection result of the acoustic wave transit time detection method based on the full convolution neural network according to the embodiment of the present invention is closer to the true value (i.e., the first arrival time of the ultrasonic signal measured by the conventional AIC method in the absence of noise) than the detection result of the conventional AIC method. Table 1 shows the average error of 16384 pieces of simulation data between the above-mentioned method for detecting acoustic wave transit time provided by the embodiment of the present invention and the conventional AIC method, and table 2 shows the standard deviation of the error of 16384 pieces of simulation data between the above-mentioned method for detecting acoustic wave transit time provided by the embodiment of the present invention and the conventional AIC method, and it can be seen from tables 1 and 2 that the above-mentioned method for detecting acoustic wave transit time provided by the embodiment of the present invention has higher detection accuracy and stability.
TABLE 1
| Signal to noise ratio snr | Mean error of the invention | Average error of AIC method |
| 37dB | 0.0599μs | 0.4463μs |
| 28dB | 0.0951μs | 0.6000μs |
TABLE 2
| Signal to noise ratio snr | Standard deviation of error of the invention | Standard deviation of error of AIC method |
| 37dB | 0.1367μs | 0.3280μs |
| 28dB | 0.2356μs | 0.5314μs |
According to the method for detecting the transit time of the sound wave, provided by the embodiment of the invention, the full convolution neural network is applied to the detection of the transit time in the medical ultrasonic CT, so that higher transit time detection precision and better noise robustness can be obtained. Compared with the traditional transit time detection method which only can combine a plurality of characteristics of the input signal, the sound wave transit time detection method based on the full convolution neural network can automatically extract a plurality of characteristics of different levels, not only can utilize the local detail information of the input signal, but also can combine the overall trend of the input signal for analysis, thereby obtaining more accurate results. In addition, the invention adopts the simulation data which is easy to mark for training, does not need to manually mark huge data sets one by one, can save the detection time and reduce the detection cost.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (3)
1. A method for detecting the transit time of sound waves based on a full convolution neural network is characterized by comprising the following steps:
s1: carrying out time window interception on a signal to be detected, keeping the length of interception consistent, recording the corresponding time of a time window starting point in the signal to be detected, and carrying out absolute value taking and normalization processing on the signal to be detected;
s2: inputting the processed signal to be detected into a trained full convolution neural network model to obtain an output function;
s3: searching the time corresponding to the maximum value of the output function in the signal to be detected, wherein the sum of the time corresponding to the maximum value of the output function in the signal to be detected and the time corresponding to the time window starting point recorded in the process of intercepting the time window of the signal to be detected in the signal to be detected is the detected transition time;
the training process of the full convolution neural network model comprises the following steps:
s21: acquiring simulation data by utilizing a sound velocity region divided by the nuclear magnetic resonance image;
s22: marking the obtained simulation data to obtain an ultrasonic signal first arrival time, wherein the shape of a marked label adopts a one-dimensional Gaussian function, and the ultrasonic signal first arrival time is taken as the central position of the Gaussian function;
s23: calculating the range of the first arrival time of the ultrasonic signals according to the distance between the ultrasonic transducers, carrying out time window interception on the simulation data and the label according to the calculated range of the first arrival time of the ultrasonic signals, keeping the length of interception consistent, recording the corresponding time of a time window starting point in the simulation data, adding white noise in the simulation data, carrying out absolute value taking and normalization processing on the simulation data added with the white noise, and finishing the manufacture of a training set;
s24: establishing a full convolution neural network model, inputting the manufactured training set into the full convolution neural network model, and training the full convolution neural network model;
s25: and continuously adjusting the weight of the full convolution neural network model by utilizing back propagation, and updating the structural parameters of the full convolution neural network model to obtain the full convolution neural network model with the optimal global parameter matrix.
2. The method for detecting acoustic wave transit time according to claim 1, wherein step S21, obtaining simulation data by using the sound velocity region divided by the nuclear magnetic resonance image, specifically includes the following steps:
s211: obtaining a plurality of nuclear magnetic resonance images, carrying out image segmentation on the nuclear magnetic resonance images, and dividing a sound velocity region according to a segmentation result;
s212: and solving a wave equation of the ultrasonic wave passing through the sound velocity region by using a finite element method to obtain simulation data.
3. The method for detecting acoustic wave transit time according to claim 1, wherein in step S22, the obtained simulation data is labeled to obtain an ultrasonic signal first arrival time, a one-dimensional gaussian function is adopted for a labeled tag shape, and the ultrasonic signal first arrival time is taken as a center position of the gaussian function, which specifically includes the following steps:
s221: and measuring the first arrival time of the ultrasonic signal by using the noiseless characteristic of the simulation signal and adopting an AIC (advanced information communication) method, wherein the shape of the labeled label adopts a one-dimensional Gaussian function, and the first arrival time of the ultrasonic signal is taken as the central position of the Gaussian function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677510.5A CN110400360B (en) | 2019-07-25 | 2019-07-25 | Sound wave transit time detection method based on full convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910677510.5A CN110400360B (en) | 2019-07-25 | 2019-07-25 | Sound wave transit time detection method based on full convolution neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110400360A CN110400360A (en) | 2019-11-01 |
| CN110400360B true CN110400360B (en) | 2021-03-19 |
Family
ID=68326025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910677510.5A Active CN110400360B (en) | 2019-07-25 | 2019-07-25 | Sound wave transit time detection method based on full convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110400360B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP4162401A1 (en) * | 2020-06-09 | 2023-04-12 | Noble International, LLC | Artificial intelligence for sound detection from medicament or test device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1664514A (en) * | 2004-03-01 | 2005-09-07 | 流速测量计有限公司 | Ultrasonic counter for determining flux of yielding medium |
| CN105596026A (en) * | 2014-06-11 | 2016-05-25 | 深圳开立生物医疗科技股份有限公司 | Method and system for ultrasound contrast imaging analysis |
Family Cites Families (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1552307A1 (en) * | 2002-10-18 | 2005-07-13 | Rescom GmbH | Diagnosis of glaucoma by complex autoantibody repertoires in body fluids |
| CN100515345C (en) * | 2003-06-12 | 2009-07-22 | 伯拉考开发股份有限公司 | Blood flow estimate through replenishment curve fitting in ultrasound contrast imaging |
| US7096135B2 (en) * | 2004-03-12 | 2006-08-22 | Panametrics, Inc. | Method and system for calculating the transit time of an ultrasonic pulse |
| US7744535B2 (en) * | 2004-07-30 | 2010-06-29 | Wisconsin Alumni Research Foundation | Method and apparatus for acoustoelastic extraction of strain and material properties |
| WO2006109203A1 (en) * | 2005-04-14 | 2006-10-19 | Koninklijke Philips Electronics N.V. | Three-dimensional time-of-flight pet with course angular and slice rebinning |
| CN101162164A (en) * | 2007-11-16 | 2008-04-16 | 浙江理工大学 | Frequency modulation wave marking method for time-difference process ultrasonic flowmeter |
| CN101790955B (en) * | 2010-01-29 | 2011-08-10 | 中国农业科学院农田灌溉研究所 | Method for controlling irrigation according to crop water shortage degree and device thereof |
| CN102359784B (en) * | 2011-08-01 | 2013-07-24 | 东北大学 | Autonomous navigation and obstacle avoidance system and method of indoor mobile robot |
| CN102645552B (en) * | 2012-04-19 | 2014-06-25 | 北京航空航天大学 | Self-adaptation related algorithm for measuring transition time |
| LU92688B1 (en) * | 2015-04-01 | 2016-10-03 | Iee Int Electronics & Eng Sa | Method and system for real-time motion artifact handling and noise removal for tof sensor images |
| US10935645B2 (en) * | 2016-10-25 | 2021-03-02 | George Mason University | Method and apparatus for low-power ultraportable ultrasound imaging |
| US10186038B1 (en) * | 2017-07-18 | 2019-01-22 | Siemens Healthcare Gmbh | Segmentation and representation network for pose discrimination |
| CN108507607B (en) * | 2018-02-09 | 2020-07-03 | 中国科学院国家空间科学中心 | Weak signal detection method based on kernel function |
| CN109784242B (en) * | 2018-12-31 | 2022-10-25 | 陕西师范大学 | Electroencephalogram signal denoising method based on one-dimensional residual convolution neural network |
| CN109875606B (en) * | 2019-01-22 | 2020-05-22 | 北京航空航天大学 | Ultrasonic CT sound velocity imaging method based on prior reflection imaging |
-
2019
- 2019-07-25 CN CN201910677510.5A patent/CN110400360B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1664514A (en) * | 2004-03-01 | 2005-09-07 | 流速测量计有限公司 | Ultrasonic counter for determining flux of yielding medium |
| CN105596026A (en) * | 2014-06-11 | 2016-05-25 | 深圳开立生物医疗科技股份有限公司 | Method and system for ultrasound contrast imaging analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110400360A (en) | 2019-11-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110353729B (en) | Sound wave transit time detection method based on bidirectional long-short term memory network | |
| KR100893286B1 (en) | Image processing system and method | |
| CN106646373B (en) | Based on way angle of arrival more than shallow sea and reach the short distance sound source joint positioning method of time delay | |
| Helble et al. | Automated acoustic localization and call association for vocalizing humpback whales on the Navy's Pacific Missile Range Facility | |
| Nosal | Methods for tracking multiple marine mammals with wide-baseline passive acoustic arrays | |
| CN109188362B (en) | Microphone array sound source positioning signal processing method | |
| CN108957403B (en) | Gaussian fitting envelope time delay estimation method and system based on generalized cross correlation | |
| CN102764141B (en) | Elastography method, elastography system, and biological tissue displacement estimation method and biological tissue displacement estimation system in elastography | |
| CN107707324A (en) | A kind of acoustical signal delay time estimation method based on phase difference and maximal possibility estimation | |
| CN114757241B (en) | Doppler parameter coupling line extraction method | |
| CN106659471A (en) | Autocorrelation guided cross-correlation in ultrasound shear wave elastography | |
| CN110554428A (en) | Seismic wave low-frequency energy change rate extraction method based on variational modal decomposition | |
| CN110400360B (en) | Sound wave transit time detection method based on full convolution neural network | |
| CN113064147A (en) | Novel matching field passive positioning method under low signal-to-noise ratio | |
| RU2550576C1 (en) | Method to measure distance to noisy object | |
| CN113063593B (en) | Rolling bearing signal characteristic information acquisition method based on ultrasonic echo energy coefficient | |
| CN109061558A (en) | A kind of sound collision detection and sound localization method based on deep learning | |
| CN106680823B (en) | Method and device for detecting target distance and speed by using sound pulse of sperm whale | |
| Sun et al. | Akte-liquid: Acoustic-based liquid identification with smartphones | |
| CN108646248A (en) | A kind of passive acoustics for low-speed motion sound source tests the speed distance measuring method | |
| CN113314127B (en) | Bird song identification method, system, computer equipment and medium based on space orientation | |
| Xu et al. | Analysis of moving source characteristics using polynomial chirplet transform | |
| Grigoriev et al. | Assessing the possibilities of identifying bowhead whales on the arctic shelf | |
| Chen et al. | A Fast Peak‐Searching Algorithm for Ultrasonic Elastography | |
| CN106580276B (en) | Pulse wave conduction time obtaining method based on correlation |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221103 Address after: Room 209-7, No. 10, Jinshan Road, High tech Zone, Suzhou City, Jiangsu Province, 215011 Patentee after: Suzhou Erxiang Foil Technology Co.,Ltd. Address before: 100191 No. 37, Haidian District, Beijing, Xueyuan Road Patentee before: BEIHANG University |