CN108056789A - A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device - Google Patents
A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device Download PDFInfo
- Publication number
- CN108056789A CN108056789A CN201711371883.7A CN201711371883A CN108056789A CN 108056789 A CN108056789 A CN 108056789A CN 201711371883 A CN201711371883 A CN 201711371883A CN 108056789 A CN108056789 A CN 108056789A
- Authority
- CN
- China
- Prior art keywords
- ultrasound scan
- scan images
- feature data
- value
- ultrasound
- 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.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/54—Control of the diagnostic device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
Abstract
The present invention provides quality preferably the first ultrasound scan images and its image quality value of a kind of same bodily tissue using several patients, and relatively low the second ultrasound scan images of quality and its image quality value, and it creates the multilayer convolutional neural networks for being capable of output image quality assessed value and the Recognition with Recurrent Neural Network of the configuration parameter list of optimization can be exported；When in use, can first initial ultrasound scan image be gathered using initial cycle neutral net, when the image quality measure value of the initial ultrasound scan image obtained if based on the multilayer convolutional neural networks is undesirable, parameter list is distributed rationally using what Recognition with Recurrent Neural Network obtained the initial ultrasound scan image, and when distributing the obtained ultrasound scan images of parameter list rationally based on this and meeting the requirements, parameter list is distributed rationally with regard to output, one has just been obtained and has preferably configured parameter list.
Description
Technical field
The present invention relates to echocardiography field more particularly to a kind of configuration parameter values for generating ultrasound scanning device
Method and apparatus.
Background technology
In practice, doctor often needs to use ultrasound scanning device and carries out ultrasonic scanning to the tissue of patient,
Process is：Ultrasonic probe emits ultrasonic wave to tissue, can occur since ultrasonic wave runs into tissue in communication process
Phenomena such as transmission, reflection and scattering, therefore, ultrasound scanning device can carry out letter according to the ultrasound echo signal received
Number detection and imaging, so as to obtain to characterize the ultrasound scan images of the structure feature of tissue, and in display
Upper display.
Generally, due to the depth of different detection positions, morphosis and tissue acoustic characteristic etc. there are some differences,
Doctor is needed according to actual conditions to several configuration parameter values of ultrasound scanning device (for example, TGC, entire gain, dynamic model
It encloses, frequency or the depth of focus etc., here, TGC's should be spelling as Time Gain Compensate, and Chinese meaning is time increasing
Benefit compensation) it is adjusted, to obtain the image of better quality, but these operations can increase Diagnostic Time, and depend on doctor
Raw subjective judgement and experience.
In the prior art, it will usually which empirically value sets one group suitable for most people in ultrasound scanning device
Configuration parameter value, so as in ultrasonic scanning, ultrasound scanning device can automatically issue this group of configuration parameter value；It can manage
Solution, since the fat or thin degree of different patients and constitution etc. there may be difference, even if so as to cause using same
Group configuration parameter value, received ultrasound echo signal can be also slightly different, so that obtained ultrasound scan images
It is of low quality.
Therefore, a kind of configuration parameter value and ultrasound scan images of getable high quality of can obtaining automatically are designed
Method just becomes a urgent problem to be solved.
The content of the invention
It is an object of the invention to provide a kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device.
In order to realize the foregoing invention first purpose, an embodiment of the present invention provides a kind of side for generating neutral net
Method comprises the following steps：
Obtain several first ultrasound scan images of the default tissue of several patients and several second ultrasonic scanning figures
Picture, the picture quality of several first ultrasound scan images are below the image matter of several second ultrasound scan images
Amount；
By the first ultrasound scan images of each patient and corresponding image quality value, the second ultrasound scan images and
Corresponding image quality value is as one group of fisrt feature data, so as to create several groups of fisrt feature data；Described image quality
It is worth and first, second ultrasound scan images is handled to obtain according to pre-set image algorithm, and the first ultrasonic scanning
The image quality value of image is respectively less than the image quality value of the second ultrasound scan images；
Default multilayer convolutional neural networks are trained based on several groups of fisrt feature data so that pass through institute
Training is stated so that the multilayer convolutional neural networks are less than first based on the quality assessment value that the first ultrasound scan images are exported
The pre-set threshold value and quality assessment value exported based on the second ultrasound scan images is more than the first pre-set threshold value.
As being further improved for an embodiment of the present invention, described image mass value is foundation pre-set image algorithm to institute
State what first, second ultrasound scan images were handled, including：Scale, circumstance of occlusion, image light according to tissue
Any one or more in illumination, image resolution ratio and picture noise situation combines, to first, second ultrasound scan images
It is handled to obtain image quality value.
It is described to be based on several groups of fisrt feature data to default as being further improved for an embodiment of the present invention
Multilayer convolutional neural networks be trained, including：Continue from untrained one group of several groups of fisrt feature data acquisitions
Fisrt feature data, and the multilayer convolutional neural networks are trained based on this group of fisrt feature data, if until described
Dry group fisrt feature data have all been trained to or the multilayer convolutional neural networks meet preset condition；
It is described that the multilayer convolutional neural networks are trained based on this group of fisrt feature data, including：It determines described
The quality assessment value that first ultrasound scan images of the multilayer convolutional neural networks based on this group of fisrt feature data are exported is more than
Equal to the first pre-set threshold value or quality assessment value that the second ultrasound scan images based on this group of fisrt feature data are exported is small
When equal to the first pre-set threshold value, the parameter of the multilayer convolutional neural networks is adjusted, until based on this group of fisrt feature data
The quality assessment value that is exported of the first ultrasound scan images be less than the first pre-set threshold value, and based on this group of fisrt feature data
The quality assessment value that is exported of the second ultrasound scan images be more than the first pre-set threshold value.
It is described to be based on several groups of fisrt feature data to default as being further improved for an embodiment of the present invention
Multilayer convolutional neural networks be trained, including：It creates one and includes input layer, multiple convolutional layers, pond layer, articulamentum
With the multilayer convolutional neural networks of output layer, and based on several groups of fisrt feature data to default multilayer convolutional Neural net
Network is trained.
It is further comprising the steps of as being further improved for an embodiment of the present invention：
The configuration parameter value list of the corresponding ultrasound scanning device of each first, second ultrasound scan images is obtained, it is described
Configuration parameter value list includes the ultrasound scanning device and each configures the corresponding configuration parameter value of parameter；
The first ultrasound scan images and corresponding configuration parameter list and the second ultrasonic scanning figure based on each patient
Picture and corresponding configuration parameter list are as one group of second feature data, so as to create several groups of second feature data；
Default Recognition with Recurrent Neural Network is trained based on several second feature data, until the Recognition with Recurrent Neural Network
Meet preset condition.
As being further improved for an embodiment of the present invention, the fisrt feature data based on several patients, second
Characteristic is trained default Recognition with Recurrent Neural Network, including：It creates one and includes input layer, several hidden layers and defeated
Go out the Recognition with Recurrent Neural Network of layer, and the fisrt feature data based on several patients, second feature data are to default Xun Huan nerve
Network is trained.
It is described to be based on several second feature data to default cycling as being further improved for an embodiment of the present invention
Neutral net is trained, until the Recognition with Recurrent Neural Network meets preset condition：Including：
Continue from the untrained second feature data of one group of several groups of second feature data acquisitions, and based on the group the
Two characteristics are trained the Recognition with Recurrent Neural Network, until several patients are trained to or the Recognition with Recurrent Neural Network
Meet preset condition；
It is described based on this group of second feature data the Recognition with Recurrent Neural Network is trained including：By this group of second feature
The default layer in the first ultrasound scan images and its corresponding configuration parameter list, the multilayer convolutional neural networks in data
Based on the intermediate data that the first ultrasound scan images in this group of second feature data are exported, for the Recognition with Recurrent Neural Network
It is trained, the default layer is multiple convolutional layers, pond layer, articulamentum or the output in the multilayer convolutional neural networks
Layer；Record the first state of several hidden layers of the Recognition with Recurrent Neural Network；By the second ultrasound in this group of second feature data
Default layer in scan image and its corresponding configuration parameter list, the multilayer convolutional neural networks is based on this group of second feature
The intermediate data that the second ultrasound scan images in data are exported is trained for the multilayer convolutional neural networks；Note
The second state of several hidden layers of the Recognition with Recurrent Neural Network is recorded, the cycling nerve net is adjusted based on first, second state
The parameter of network.
It is described until the Recognition with Recurrent Neural Network meets default item as being further improved for an embodiment of the present invention
Part, including：
Reach preset value or the Recognition with Recurrent Neural Network in the quantity for determining housebroken second feature data and be based on the
The output of the first ultrasound scan images in two characteristics is estimated in configuration parameter list and second feature data
When the difference of the configuration parameter list of second ultrasound scan images is less than the second pre-set threshold value, meet preset condition.
An embodiment of the present invention provides a kind of device for generating neutral net, including with lower module：
First parameter acquisition module, for obtaining several first ultrasound scan images of the default tissue of several patients
With several second ultrasound scan images, the picture quality of several first ultrasound scan images is below described several the second surpass
The picture quality of sound scan image；
Characteristic generation module, for by the first ultrasound scan images of each patient and corresponding picture quality
Value, the second ultrasound scan images and corresponding image quality value are as one group of fisrt feature data, so as to create several groups the
One characteristic；Described image mass value be according to pre-set image algorithm to first, second ultrasound scan images at
What reason obtained, and the image quality value of the first ultrasound scan images is respectively less than the image quality value of the second ultrasound scan images；
Neutral net generation module, for being based on several groups of fisrt feature data to default multilayer convolutional Neural net
Network is trained so that is exported by the training multilayer convolutional neural networks based on the first ultrasound scan images
The quality assessment value quality assessment value that is less than the first pre-set threshold value and is exported based on the second ultrasound scan images be more than first
Pre-set threshold value.
An embodiment of the present invention provides a kind of method of the generation configuration parameter list for ultrasound scanning device, bag
Include following steps：
Initial configuration parameters list is obtained, and the ultrasound echo signal received is carried out based on initial configuration parameters list
Detection and imaging, obtain initial ultrasound scan image；
Based on the multilayer convolutional neural networks obtain the initial ultrasound scan image quality assessment value do not meet will
When asking, initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing parameter list rationally；
The ultrasound echo signal received is detected and imaging, is optimized based on parameter list is distributed rationally
Ultrasound scan images；It determines to obtain the quality assessment value of the optimization ultrasound scan images based on the multilayer convolutional neural networks
When meeting the requirements, then it is the configuration parameter list to distribute parameter list rationally.
It is further comprising the steps of as being further improved for an embodiment of the present invention：
The quality assessment value for determining to be obtained the optimization ultrasound scan images based on the multilayer convolutional neural networks is met
When not requiring, the parameter for adjusting the Recognition with Recurrent Neural Network is continuously carried out, and initial ultrasound scan image is inputted into the cycling
Obtain distributing rationally parameter list in neutral net, and based on distribute rationally parameter list to the ultrasound echo signal that is received into
Row detection and imaging obtain optimization ultrasound scan images, until determining to obtain institute based on the multilayer convolutional neural networks
Quality assessment value meet the requirements or the adjust number of parameter of the Recognition with Recurrent Neural Network for stating optimization ultrasound scan images is more than
Preset times.
An embodiment of the present invention provides a kind of device of the generation configuration parameter list for ultrasound scanning device, bag
It includes with lower module：
First parameter acquisition module, for obtaining initial configuration parameters list, and based on initial configuration parameters list to institute
The ultrasound echo signal of reception is detected and imaging, obtains initial ultrasound scan image；
Quality assessment modules, for obtaining the matter of the initial ultrasound scan image based on the multilayer convolutional neural networks
When amount assessed value is undesirable, initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing ginseng rationally
Ordered series of numbers table；
Judgment module, for the ultrasound echo signal received being detected and being imaged based on distributing parameter list rationally
Processing obtains optimization ultrasound scan images；It determines to obtain the optimization ultrasonic scanning figure based on the multilayer convolutional neural networks
When the quality assessment value of picture meets the requirements, then it is the configuration parameter list to distribute parameter list rationally.
Compared with the prior art, the technical effects of the invention are that：Present invention offer is a kind of to utilize the same of several patients
The quality of bodily tissue preferably the first ultrasound scan images and its relatively low the second ultrasonic scanning of image quality value and quality
Image and its image quality value, and create the multilayer convolutional neural networks for being capable of output image quality assessed value, Yi Jineng
The Recognition with Recurrent Neural Network of the configuration parameter list of enough output optimization；When in use, can first be adopted using initial cycle neutral net
Collect initial ultrasound scan image, if based on the image for the initial ultrasound scan image that the multilayer convolutional neural networks obtain
When quality assessment value is undesirable, the parameter of distributing rationally that the initial ultrasound scan image is obtained using Recognition with Recurrent Neural Network is arranged
Table, and when distributing the obtained ultrasound scan images of parameter list rationally based on this and meeting the requirements rationally, parameter is distributed with regard to output
List has just obtained one and has preferably configured parameter list.
Description of the drawings
Fig. 1 is the first flow diagram of the method for the generation neutral net in the embodiment of the present invention one；
Fig. 2 is second of flow diagram of the method for the generation neutral net in the embodiment of the present invention one；
Fig. 3 is the third flow diagram of the method for the generation neutral net in the embodiment of the present invention one；
Fig. 4 is the first flow diagram of the method for the generation configuration parameter list in the embodiment of the present invention one；
Fig. 5 is second of flow diagram of the method for the generation configuration parameter list in the embodiment of the present invention one.
Specific embodiment
Below with reference to each embodiment shown in the drawings, the present invention will be described in detail.But these embodiments are not
The limitation present invention, structure that those of ordinary skill in the art are made according to these embodiments, method or change functionally
It changes and is all contained in protection scope of the present invention.
The embodiment of the present invention one provides a kind of method for generating neutral net, here, the method for the generation neutral net
It can be performed by the control system in ultrasound scanning device, as shown in Figure 1, comprising the following steps：
Step 101：It obtains several first ultrasound scan images of the default tissue of several patients and several the second surpasses
Sound scan image, the picture quality of several first ultrasound scan images are below several second ultrasound scan images
Picture quality；Here, since ultrasound scanning device is provided with several configuration parameters, in order to enable ultrasound scanning device energy
Enough normal works to each configuration parameter, it is necessary to set configuration parameter value, and can use a list (i.e. configuration parameter value
List) store these configuration parameter values；It is understood that the configuration parameter value list can be divided into two classes：(1) it is suitable for
The configuration parameter value list of most people, i.e., no optimised mistake, at this point, ultrasound scanning device can just collect the first ultrasound
Scan image；(2) the configuration parameter value list optimized by doctor or expert, at this point, ultrasound scanning device can just collect
Second ultrasound scan images.Here, the first scan image and the second scanning can be gathered to the default tissue of each patient
Image, it is possible to understand that be that the quantity of several patients is equal to the quantity of the first scan image, be also equal to the number of the second scan image
Amount.
Step 102：By the first ultrasound scan images of each patient and corresponding image quality value, the second ultrasonic scanning
Image and corresponding image quality value are as one group of fisrt feature data, so as to create several groups of fisrt feature data；It is described
Image quality value is handled to obtain according to pre-set image algorithm to first, second ultrasound scan images, and first
The image quality value of ultrasound scan images is respectively less than the image quality value of the second ultrasound scan images；Fisrt feature data include two
Type：(1) first ultrasound scan images and corresponding image quality value；(2) second ultrasound scan images and corresponding figure
Image quality magnitude.And be appreciated that, the quantitative values of several groups of fisrt feature data is equal to the first ultrasound scan images and the
The sum of quantitative value of two ultrasound scan images.In this step, first, second ultrasound scan images can be realized using program
To the mapping of image quality value, i.e. each first, second ultrasound scan images for input, program can be exported and be used for
Characterize the image quality value of its picture quality.Optionally, the image quality value of first, second ultrasound scan images can be by artificial
Input.
Step 103：Default multilayer convolutional neural networks are trained based on several groups of fisrt feature data, are made
The quality assessment value that must be exported by the training multilayer convolutional neural networks based on the first ultrasound scan images
The quality assessment value exported less than the first pre-set threshold value and based on the second ultrasound scan images is more than the first pre-set threshold value.
In the step, a multilayer convolutional neural networks can be pre-created, then using the fisrt feature data of several patients and
Two characteristics are trained the multilayer convolutional neural networks, can be constantly to multilayer convolution god during the training
(for example, being finely adjusted to constant in excitation function etc.) is finely adjusted through network, until the multilayer convolutional neural networks are based on
The quality assessment value that first ultrasound scan images are exported is less than the first pre-set threshold value and defeated based on the second ultrasound scan images institute
The quality assessment value gone out is more than the first pre-set threshold value.
Here, the multilayer convolutional Neural net has been trained in the fisrt feature data and second feature data for using several patients
After network, which can export quality assessment value to the ultrasound scan images of input, it is to be understood that
If quality assessment value is more than the first pre-set threshold value, the quality of the ultrasound scan images is more excellent；Otherwise, quality is with regard to relatively low.
Preferably, described image mass value be according to pre-set image algorithm to first, second ultrasound scan images into
Row processing obtains, including：Scale, circumstance of occlusion, image light illumination, image resolution ratio and picture noise according to tissue
Any one or more in situation combines, and first, second ultrasound scan images are handled to obtain image quality value.This
In, scale, circumstance of occlusion, image light illumination, image resolution ratio and the picture noise situation of tissue accurate can be retouched
The picture quality of ultrasound scan images is stated, therefore, the figure of first, second ultrasound scan images can be assessed using these indexs
Image quality amount.
Preferably, it is described that default multilayer convolutional neural networks are instructed based on several groups of fisrt feature data
Practice, including：
Continue from the untrained one group of fisrt feature data of several groups of fisrt feature data acquisitions, and based on the group the
One characteristic is trained the multilayer convolutional neural networks, until several groups of fisrt feature data are all instructed
Experienced or described multilayer convolutional neural networks meet preset condition；
It is described that the multilayer convolutional neural networks are trained based on this group of fisrt feature data, including：It determines described
The quality assessment value that first ultrasound scan images of the multilayer convolutional neural networks based on this group of fisrt feature data are exported is more than
Equal to the first pre-set threshold value or quality assessment value that the second ultrasound scan images based on this group of fisrt feature data are exported is small
When equal to the first pre-set threshold value, the parameter of the multilayer convolutional neural networks is adjusted, until based on this group of fisrt feature data
The quality assessment value that is exported of the first ultrasound scan images be less than the first pre-set threshold value, and based on this group of fisrt feature data
The quality assessment value that is exported of the second ultrasound scan images be more than the first pre-set threshold value.
It here, can not during several groups of fisrt feature data is used to be trained multilayer convolutional neural networks
The parameter of disconnected adjustment multilayer convolutional neural networks, until the output valve to multilayer convolutional neural networks is met the requirements, i.e., first
Ultrasound scan images and the quality assessment value that exports is less than the first pre-set threshold value and the second ultrasound scan images and the quality that exports
Assessed value is more than the first pre-set threshold value.
As shown in Fig. 2, this training process may comprise steps of：
Step 201：It creates default multilayer convolutional neural networks and is initialized；
Step 202：Obtain first, second ultrasound scan images of several patients and its corresponding quality assessment value；
Step 203：First, second ultrasound scan images of one untrained patient of acquisition and corresponding picture quality
Value；
Step 204：First, second ultrasound scan images to be trained and corresponding matter are inputted to multilayer convolutional neural networks
Assessed value is measured, and obtains quality assessment value；
Step 205：Judge whether quality assessment value meets the requirementsWhether quality assessment value during the first ultrasound scan images
Less than the first pre-set threshold valueAnd whether the quality assessment value of the second ultrasound scan images is more than the first pre-set threshold value；Then meet
It is required that operating procedure 206 afterwards；Otherwise it is unsatisfactory for requiring, afterwards operating procedure 207；
Step 206：Whether all trained complete or learning efficiency of first, second ultrasound scan images of all patients has expired
Sufficient conditionIf it is not, then performing step 203, step 208 is otherwise performed；
Step 207：The parameter of the multilayer convolutional neural networks is adjusted, and performs step 204；
Step 208：Training is completed.
Preferably, it is described that default multilayer convolutional neural networks are instructed based on several groups of fisrt feature data
Practice, including：It creates one and includes input layer, multiple convolutional layers, pond layer, the multilayer convolutional Neural net of articulamentum and output layer
Network, and default multilayer convolutional neural networks are trained based on several groups of fisrt feature data.
Preferably, it is further comprising the steps of：
The configuration parameter value list of the corresponding ultrasound scanning device of each first, second ultrasound scan images is obtained, it is described
Configuration parameter value list includes the ultrasound scanning device and each configures the corresponding configuration parameter value of parameter；
The first ultrasound scan images and corresponding configuration parameter list and the second ultrasonic scanning figure based on each patient
Picture and corresponding configuration parameter list are as one group of second feature data, so as to create several groups of second feature data；
Default Recognition with Recurrent Neural Network is trained based on several second feature data, until the Recognition with Recurrent Neural Network
Meet preset condition.
In this step, a Recognition with Recurrent Neural Network can be pre-created, then using several groups of second feature data pair
The Recognition with Recurrent Neural Network is trained, constantly the Recognition with Recurrent Neural Network can be finely adjusted during the training (for example,
Constant in excitation function etc. is finely adjusted), until the Recognition with Recurrent Neural Network meets preset condition.
Here, after several groups of second feature data is used to train the Recognition with Recurrent Neural Network, the Recognition with Recurrent Neural Network
A more preferably configuration parameter value list can be exported to the ultrasound scan images of input, and if it is understood that used
The configuration parameter value list configuration ultrasound scanning device, then when scanning again, which can be adopted with larger probability
Collect quality more preferably ultrasound scan images.
Preferably, the fisrt feature data based on several patients, second feature data are to default cycling nerve net
Network is trained, including：It creates one and includes the Recognition with Recurrent Neural Network of input layer, several hidden layers and output layer, and be based on
Fisrt feature data, the second feature data of several patients are trained default Recognition with Recurrent Neural Network.
Preferably, it is described that default Recognition with Recurrent Neural Network is trained based on several second feature data, until described
Recognition with Recurrent Neural Network meets preset condition, including：Continue from one group of several groups of second feature data acquisitions untrained the
Two characteristics, and the Recognition with Recurrent Neural Network is trained based on this group of second feature data, until several patients
It is trained to or the Recognition with Recurrent Neural Network meets preset condition；
It is described based on this group of second feature data the Recognition with Recurrent Neural Network is trained including：
By the first ultrasound scan images in this group of second feature data and its corresponding configuration parameter list, the multilayer
The mediant that default layer in convolutional neural networks is exported based on the first ultrasound scan images in this group of second feature data
According to, be trained for the Recognition with Recurrent Neural Network, the default layer be the multilayer convolutional neural networks in multiple convolution
Layer, pond layer, articulamentum or output layer；
Record the first state of several hidden layers of the Recognition with Recurrent Neural Network；
By the second ultrasound scan images in this group of second feature data and its corresponding configuration parameter list, the multilayer
The mediant that default layer in convolutional neural networks is exported based on the second ultrasound scan images in this group of second feature data
According to being trained for the multilayer convolutional neural networks；
The second state of several hidden layers of the Recognition with Recurrent Neural Network is recorded, based on described in the adjustment of first, second state
The parameter of Recognition with Recurrent Neural Network.
Preferably, it is described until the Recognition with Recurrent Neural Network meet preset condition, including：Determining housebroken second
The quantity of characteristic reaches preset value or the Recognition with Recurrent Neural Network based on the first ultrasonic scanning figure in second feature data
The configuration parameter row for estimating configuration parameter list and the second ultrasound scan images in second feature data of the output of picture
When the difference of table is less than the second pre-set threshold value, meet preset condition.
As shown in figure 3, this training process may comprise steps of：
Step 301：It creates default Recognition with Recurrent Neural Network and is initialized
Step 302：First, second ultrasound scan images of several patients and its corresponding configuration parameter list are obtained, and
Form several groups of second feature data；
Step 303：From the untrained second feature data of one group of several groups of second feature data acquisitions；
Step 304：By the first ultrasound scan images in this group of second feature data and its it is corresponding configuration parameter list,
Default layer in the multilayer convolutional neural networks is exported based on the first ultrasound scan images in this group of second feature data
Intermediate data, be trained for the Recognition with Recurrent Neural Network；
Step 305：Record the first state of several hidden layers of the Recognition with Recurrent Neural Network；
Step 306：By the second ultrasound scan images in this group of second feature data and its it is corresponding configuration parameter list,
Default layer in the multilayer convolutional neural networks is exported based on the second ultrasound scan images in this group of second feature data
Intermediate data, be trained for the multilayer convolutional neural networks；
Step 307：The second state of several hidden layers of the Recognition with Recurrent Neural Network is recorded, based on first, second state
Adjust the parameter of the Recognition with Recurrent Neural Network；
Step 308：Judge that the Recognition with Recurrent Neural Network meets preset conditionI.e. in definite housebroken second feature number
According to quantity reach the institute of preset value or the Recognition with Recurrent Neural Network based on the first ultrasound scan images in second feature data
State the difference for estimating configuration parameter list and the configuration parameter list of the second ultrasound scan images in second feature data of output
It is different when being less than the second pre-set threshold value, meet preset condition；Step 309 is performed if meeting, otherwise performs step 303；
Step 309：Deconditioning.
An embodiment of the present invention provides a kind of device for generating neutral net, including with lower module：
Parameter acquisition module, if for obtain several first ultrasound scan images of the default tissue of several patients and
Dry second ultrasound scan images, the picture quality of several first ultrasound scan images are below several second ultrasounds and sweep
The picture quality of tracing picture；
Characteristic generation module, for by the first ultrasound scan images of each patient and corresponding picture quality
Value, the second ultrasound scan images and corresponding image quality value are as one group of fisrt feature data, so as to create several groups the
One characteristic；Described image mass value be according to pre-set image algorithm to first, second ultrasound scan images at
What reason obtained, and the image quality value of the first ultrasound scan images is respectively less than the image quality value of the second ultrasound scan images；
Neutral net generation module, for being based on several groups of fisrt feature data to default multilayer convolutional Neural net
Network is trained so that is exported by the training multilayer convolutional neural networks based on the first ultrasound scan images
The quality assessment value quality assessment value that is less than the first pre-set threshold value and is exported based on the second ultrasound scan images be more than first
Pre-set threshold value.
The embodiment of the present invention two provides a kind of method of the generation configuration parameter list for ultrasound scanning device, such as schemes
Shown in 5, comprise the following steps：
Step 501：Initial configuration parameters list is obtained, and based on initial configuration parameters list to the ultrasonic echo that is received
Signal is detected and imaging, obtains initial ultrasound scan image；
Step 502：The quality assessment value of the initial ultrasound scan image is obtained based on the multilayer convolutional neural networks
When undesirable, initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing parameter list rationally；
Step 503：The ultrasound echo signal received is detected and imaging based on parameter list is distributed rationally,
Obtain optimization ultrasound scan images；It determines to obtain the matter of the optimization ultrasound scan images based on the multilayer convolutional neural networks
When amount assessed value meets the requirements, then it is the configuration parameter list to distribute parameter list rationally.
Preferably, it is further comprising the steps of：
The quality assessment value for determining to be obtained the optimization ultrasound scan images based on the multilayer convolutional neural networks is met
When not requiring, the parameter for adjusting the Recognition with Recurrent Neural Network is continuously carried out, and initial ultrasound scan image is inputted into the cycling
Obtain distributing rationally parameter list in neutral net, and based on distribute rationally parameter list to the ultrasound echo signal that is received into
Row detection and imaging obtain optimization ultrasound scan images, until determining to obtain institute based on the multilayer convolutional neural networks
Quality assessment value meet the requirements or the adjust number of parameter of the Recognition with Recurrent Neural Network for stating optimization ultrasound scan images is more than
Preset times.
As shown in figure 4, the method for generation configuration parameter list may comprise steps of：
Step 401：According to the tissue that user selects, corresponding trained multilayer convolutional neural networks in advance are loaded
And Recognition with Recurrent Neural Network and default configuration parameter list；
Step 402：Emit ultrasonic wave to tissue, receive ultrasound echo signal；
Step 403：Ultrasound echo signal is detected and is imaged with default configuration parameter list, forms initial ultrasound
Scan image；
Step 404：Initial ultrasound scan image is inputted into multilayer convolutional neural networks, obtains the first quality assessment value；
Step 405：Judge that the first quality assessment value meets the requirementsIf it is, performing step 411, step is otherwise performed
406；
Step 406：The initial ultrasound image is inputted into Recognition with Recurrent Neural Network, and obtains and distributes parameter list rationally；
Step 407：Emit ultrasonic wave to tissue, ultrasound echo signal is received, with the configuration parameter list pair of optimization
Echo-signal is detected and is imaged, and forms initial ultrasound image；
Step 408：Second quality assessment value of the ultrasonoscopy after being optimized using pre-set image algorithm；
Step 409：Judge that the second quality assessment value meets to be expected, and always adjust number or take and be less than preset valueIf
It is then to perform step 411, otherwise, performs step 410；
Step 410：Recognition with Recurrent Neural Network is finely tuned, and performs step 406；
Step 411：Parameter list will be distributed rationally to show.So as to which doctor can select to distribute parameter rationally using this
List.
The embodiment of the present invention additionally provides a kind of device of the generation configuration parameter list for ultrasound scanning device, including
With lower module：
First parameter acquisition module, for obtaining initial configuration parameters list, and based on initial configuration parameters list to institute
The ultrasound echo signal of reception is detected and imaging, obtains initial ultrasound scan image；
Quality assessment modules, for obtaining the matter of the initial ultrasound scan image based on the multilayer convolutional neural networks
When amount assessed value is undesirable, initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing ginseng rationally
Ordered series of numbers table；
Judgment module, for the ultrasound echo signal received being detected and being imaged based on distributing parameter list rationally
Processing obtains optimization ultrasound scan images；It determines to obtain the optimization ultrasonic scanning figure based on the multilayer convolutional neural networks
When the quality assessment value of picture meets the requirements, then it is the configuration parameter list to distribute parameter list rationally.
It should be appreciated that although this specification is described in terms of embodiments, but not each embodiment only includes one
A independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should will say
For bright book as an entirety, the technical solution in each embodiment may also be suitably combined to form those skilled in the art can
With the other embodiment of understanding.
Those listed above is a series of to be described in detail only for feasibility embodiment of the invention specifically
Bright, they are not to limit the scope of the invention, all equivalent implementations made without departing from skill spirit of the present invention
Or change should all be included in the protection scope of the present invention.
Claims (12)
- A kind of 1. method for generating neutral net, which is characterized in that comprise the following steps：Obtain several first ultrasound scan images of the default tissue of several patients and several second ultrasound scan images, institute The picture quality for stating several first ultrasound scan images is below the picture quality of several second ultrasound scan images；By the first ultrasound scan images and corresponding image quality value, the second ultrasound scan images and correspondence of each patient Image quality value as one group of fisrt feature data, so as to create several groups of fisrt feature data；Described image mass value is First, second ultrasound scan images are handled according to pre-set image algorithm, and the first ultrasound scan images Image quality value be respectively less than the image quality values of the second ultrasound scan images；Default multilayer convolutional neural networks are trained based on several groups of fisrt feature data so that pass through the instruction Practice so that the multilayer convolutional neural networks are default less than first based on the quality assessment value that the first ultrasound scan images are exported The threshold values and quality assessment value exported based on the second ultrasound scan images is more than the first pre-set threshold value.
- 2. the method for generation neutral net according to claim 1, which is characterized in that described image mass value is according to pre- If image algorithm handles first, second ultrasound scan images, including：According to any one in the scale of tissue, circumstance of occlusion, image light illumination, image resolution ratio and picture noise situation Or multiple combinations, first, second ultrasound scan images are handled to obtain image quality value.
- 3. the method for generation neutral net according to claim 1, which is characterized in that described to be based on described several groups first Characteristic is trained default multilayer convolutional neural networks, including：Continue from the untrained one group of fisrt feature data of several groups of fisrt feature data acquisitions, and it is special based on the group first Sign data the multilayer convolutional neural networks are trained, until several groups of fisrt feature data be all trained to or The multilayer convolutional neural networks meet preset condition；It is described that the multilayer convolutional neural networks are trained based on this group of fisrt feature data, including：Determine the multilayer The quality assessment value that first ultrasound scan images of the convolutional neural networks based on this group of fisrt feature data are exported is more than or equal to The quality assessment value that first pre-set threshold value or the second ultrasound scan images based on this group of fisrt feature data are exported is less than etc. When the first pre-set threshold value, the parameter of the multilayer convolutional neural networks is adjusted, until the based on this group of fisrt feature data The quality assessment value that one ultrasound scan images are exported is less than the first pre-set threshold value, and the based on this group of fisrt feature data The quality assessment value that two ultrasound scan images are exported is more than the first pre-set threshold value.
- 4. the method for generation neutral net according to claim 1, which is characterized in that described to be based on described several groups first Characteristic is trained default multilayer convolutional neural networks, including：It creates one and includes input layer, multiple convolution The multilayer convolutional neural networks of layer, pond layer, articulamentum and output layer, and based on several groups of fisrt feature data to default Multilayer convolutional neural networks be trained.
- 5. the method for generation neutral net according to claim 1, which is characterized in that further comprising the steps of：Obtain the configuration parameter value list of the corresponding ultrasound scanning device of each first, second ultrasound scan images, the configuration Parameter value list includes the ultrasound scanning device and each configures the corresponding configuration parameter value of parameter；The first ultrasound scan images based on each patient and corresponding configuration parameter list and the second ultrasound scan images with And configuration parameter list is corresponded to as one group of second feature data, so as to create several groups of second feature data；Default Recognition with Recurrent Neural Network is trained based on several second feature data, until the Recognition with Recurrent Neural Network meets Preset condition.
- 6. it is according to claim 5 generation neutral net method, which is characterized in that it is described based on several patients first Characteristic, second feature data are trained default Recognition with Recurrent Neural Network, including：Create one include input layer, The Recognition with Recurrent Neural Network of several hidden layers and output layer, and the fisrt feature data based on several patients, second feature data pair Default Recognition with Recurrent Neural Network is trained.
- 7. the method for the generation neutral net according to right will go 6, which is characterized in that described to be based on several second feature numbers It is trained according to default Recognition with Recurrent Neural Network, until the Recognition with Recurrent Neural Network meets preset condition：Including：Continue from the untrained second feature data of one group of several groups of second feature data acquisitions, and it is special based on the group second Sign data are trained the Recognition with Recurrent Neural Network, until several patients are trained to or the Recognition with Recurrent Neural Network meets Preset condition；It is described based on this group of second feature data the Recognition with Recurrent Neural Network is trained including：By this group of second feature data In the first ultrasound scan images and its corresponding configuration parameter list, the default layer in the multilayer convolutional neural networks be based on The intermediate data that the first ultrasound scan images in this group of second feature data are exported is carried out for the Recognition with Recurrent Neural Network Training, the default layer are multiple convolutional layers, pond layer, articulamentum or the output layer in the multilayer convolutional neural networks；Note Record the first state of several hidden layers of the Recognition with Recurrent Neural Network；By the second ultrasonic scanning figure in this group of second feature data Default layer in picture and its corresponding configuration parameter list, the multilayer convolutional neural networks is based in this group of second feature data The intermediate data that is exported of the second ultrasound scan images, be trained for the multilayer convolutional neural networks；Described in record Second state of several hidden layers of Recognition with Recurrent Neural Network adjusts the ginseng of the Recognition with Recurrent Neural Network based on first, second state Number.
- 8. the method for generation neutral net according to claim 7, which is characterized in that described until the cycling nerve net Network meets preset condition, including：Reach preset value in the quantity for determining housebroken second feature data or the Recognition with Recurrent Neural Network is based on the second spy Configuration parameter list and second in second feature data are estimated in the output of the first ultrasound scan images in sign data When the difference of the configuration parameter list of ultrasound scan images is less than the second pre-set threshold value, meet preset condition.
- 9. a kind of device for generating neutral net, which is characterized in that including with lower module：First parameter acquisition module, if for obtain several first ultrasound scan images of the default tissue of several patients and Dry second ultrasound scan images, the picture quality of several first ultrasound scan images are below several second ultrasounds and sweep The picture quality of tracing picture；Characteristic generation module, for by the first ultrasound scan images of each patient and corresponding image quality value, Two ultrasound scan images and corresponding image quality value are as one group of fisrt feature data, so as to create several groups of fisrt feature Data；Described image mass value is that first, second ultrasound scan images are handled to obtain according to pre-set image algorithm , and the image quality value of the first ultrasound scan images is respectively less than the image quality value of the second ultrasound scan images；Neutral net generation module, for be based on several groups of fisrt feature data to default multilayer convolutional neural networks into Row training so that the matter exported by the training multilayer convolutional neural networks based on the first ultrasound scan images It is default that the quality assessment value that amount assessed value is less than the first pre-set threshold value and is exported based on the second ultrasound scan images is more than first Threshold values.
- 10. a kind of method of generation configuration parameter list for ultrasound scanning device, which is characterized in that comprise the following steps：Initial configuration parameters list is obtained, and the ultrasound echo signal received is detected based on initial configuration parameters list And imaging, obtain initial ultrasound scan image；Based on the multilayer convolutional neural networks obtain the initial ultrasound scan image quality assessment value it is undesirable when, Initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing parameter list rationally；The ultrasound echo signal received is detected and imaging based on parameter list is distributed rationally, obtains optimization ultrasound Scan image；The quality assessment value for determining to be obtained the optimization ultrasound scan images based on the multilayer convolutional neural networks is met It is required that when, then it is the configuration parameter list to distribute parameter list rationally.
- 11. the method for generation configuration parameter list according to claim 10, which is characterized in that further comprising the steps of：Determine to obtain the quality assessment values of the optimization ultrasound scan images based on the multilayer convolutional neural networks that meet should not When asking, the parameter for adjusting the Recognition with Recurrent Neural Network is continuously carried out, and the initial ultrasound scan image input Xun Huan is neural It obtains distributing parameter list rationally in network, and the ultrasound echo signal received is examined based on parameter list is distributed rationally Survey and imaging obtain optimization ultrasound scan images, described excellent until determining to obtain based on the multilayer convolutional neural networks It is more than default to change quality assessment value meet the requirements or the adjust number of parameter of the Recognition with Recurrent Neural Network of ultrasound scan images Number.
- 12. the device of a kind of generation configuration parameter list for ultrasound scanning device, which is characterized in that including with lower module：First parameter acquisition module, for obtaining initial configuration parameters list, and based on initial configuration parameters list to being received Ultrasound echo signal be detected and imaging, obtain initial ultrasound scan image；Quality assessment modules, the quality for being obtained the initial ultrasound scan image based on the multilayer convolutional neural networks are commented When valuation is undesirable, initial ultrasound scan image is inputted in the Recognition with Recurrent Neural Network, obtains distributing parameter row rationally Table；Judgment module, for based on distribute rationally parameter list the ultrasound echo signal received is detected and imaging at Reason obtains optimization ultrasound scan images；It determines to obtain the optimization ultrasound scan images based on the multilayer convolutional neural networks Quality assessment value when meeting the requirements, then it is the configuration parameter list to distribute parameter list rationally.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711371883.7A CN108056789A (en) | 2017-12-19 | 2017-12-19 | A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711371883.7A CN108056789A (en) | 2017-12-19 | 2017-12-19 | A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108056789A true CN108056789A (en) | 2018-05-22 |
Family
ID=62139240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711371883.7A Pending CN108056789A (en) | 2017-12-19 | 2017-12-19 | A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108056789A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109191442A (en) * | 2018-08-28 | 2019-01-11 | 深圳大学 | Ultrasound image assessment and screening technique and device |
| CN109350100A (en) * | 2018-09-27 | 2019-02-19 | 上海联影医疗科技有限公司 | Medical imaging procedure, medical imaging devices and computer readable storage medium |
| CN109976153A (en) * | 2019-03-01 | 2019-07-05 | 北京三快在线科技有限公司 | Control the method, apparatus and electronic equipment of unmanned equipment and model training |
| CN110840482A (en) * | 2019-10-28 | 2020-02-28 | 苏州佳世达电通有限公司 | Ultrasonic imaging system and method thereof |
| CN110960262A (en) * | 2019-12-31 | 2020-04-07 | 上海杏脉信息科技有限公司 | Ultrasonic scanning system, method and medium |
| CN111060591A (en) * | 2019-12-06 | 2020-04-24 | 北京瑞莱智慧科技有限公司 | Metal part fatigue monitoring method and system based on cavity convolution network |
| CN111513754A (en) * | 2019-09-16 | 2020-08-11 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging equipment and quality evaluation method of ultrasonic image |
| CN111789635A (en) * | 2019-04-04 | 2020-10-20 | 株式会社日立制作所 | Ultrasonic imaging apparatus and image processing apparatus |
| CN112890853A (en) * | 2019-12-04 | 2021-06-04 | 通用电气精准医疗有限责任公司 | System and method for joint scan parameter selection |
| CN112890854A (en) * | 2019-12-04 | 2021-06-04 | 通用电气精准医疗有限责任公司 | System and method for sequential scan parameter selection |
| JPWO2020003990A1 (en) * | 2018-06-28 | 2021-07-08 | 富士フイルム株式会社 | Medical image processing equipment and methods, machine learning systems, programs and storage media |
| CN113438451A (en) * | 2021-06-21 | 2021-09-24 | 易成功(厦门)信息科技有限公司 | Unified standardization processing platform and method for multi-terminal multi-source data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1934589A (en) * | 2004-03-23 | 2007-03-21 | 美国西门子医疗解决公司 | Systems and methods providing automated decision support for medical imaging |
| CN103222879A (en) * | 2012-01-25 | 2013-07-31 | 通用电气公司 | System and method for identifying an optimal image frame for ultrasound imaging |
| CN103237499A (en) * | 2010-04-07 | 2013-08-07 | 富士胶片索诺声公司 | Systems and methods for enhanced imaging of objects within an image |
| CN103845081A (en) * | 2012-11-28 | 2014-06-11 | 深圳迈瑞生物医疗电子股份有限公司 | System and method for ultrasonic elastography and method for real-time dynamic interframe processing |
| CN104684452A (en) * | 2012-12-26 | 2015-06-03 | 奥林巴斯医疗株式会社 | Image recording device and image recording method |
| CN106408566A (en) * | 2016-11-10 | 2017-02-15 | 深圳大学 | Fetal ultrasound image quality control method and system |
-
2017
- 2017-12-19 CN CN201711371883.7A patent/CN108056789A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1934589A (en) * | 2004-03-23 | 2007-03-21 | 美国西门子医疗解决公司 | Systems and methods providing automated decision support for medical imaging |
| CN103237499A (en) * | 2010-04-07 | 2013-08-07 | 富士胶片索诺声公司 | Systems and methods for enhanced imaging of objects within an image |
| CN103222879A (en) * | 2012-01-25 | 2013-07-31 | 通用电气公司 | System and method for identifying an optimal image frame for ultrasound imaging |
| CN103845081A (en) * | 2012-11-28 | 2014-06-11 | 深圳迈瑞生物医疗电子股份有限公司 | System and method for ultrasonic elastography and method for real-time dynamic interframe processing |
| CN104684452A (en) * | 2012-12-26 | 2015-06-03 | 奥林巴斯医疗株式会社 | Image recording device and image recording method |
| CN106408566A (en) * | 2016-11-10 | 2017-02-15 | 深圳大学 | Fetal ultrasound image quality control method and system |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7033202B2 (en) | 2018-06-28 | 2022-03-09 | 富士フイルム株式会社 | Medical image processing equipment and methods, machine learning systems, programs and storage media |
| JPWO2020003990A1 (en) * | 2018-06-28 | 2021-07-08 | 富士フイルム株式会社 | Medical image processing equipment and methods, machine learning systems, programs and storage media |
| EP3815610A4 (en) * | 2018-06-28 | 2021-09-15 | FUJIFILM Corporation | Medical-image processing device and method, machine learning system, program, and storage medium |
| CN109191442A (en) * | 2018-08-28 | 2019-01-11 | 深圳大学 | Ultrasound image assessment and screening technique and device |
| CN109350100A (en) * | 2018-09-27 | 2019-02-19 | 上海联影医疗科技有限公司 | Medical imaging procedure, medical imaging devices and computer readable storage medium |
| CN109976153A (en) * | 2019-03-01 | 2019-07-05 | 北京三快在线科技有限公司 | Control the method, apparatus and electronic equipment of unmanned equipment and model training |
| CN111789635A (en) * | 2019-04-04 | 2020-10-20 | 株式会社日立制作所 | Ultrasonic imaging apparatus and image processing apparatus |
| CN111513754A (en) * | 2019-09-16 | 2020-08-11 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging equipment and quality evaluation method of ultrasonic image |
| CN110840482A (en) * | 2019-10-28 | 2020-02-28 | 苏州佳世达电通有限公司 | Ultrasonic imaging system and method thereof |
| CN112890853A (en) * | 2019-12-04 | 2021-06-04 | 通用电气精准医疗有限责任公司 | System and method for joint scan parameter selection |
| CN112890854A (en) * | 2019-12-04 | 2021-06-04 | 通用电气精准医疗有限责任公司 | System and method for sequential scan parameter selection |
| CN111060591A (en) * | 2019-12-06 | 2020-04-24 | 北京瑞莱智慧科技有限公司 | Metal part fatigue monitoring method and system based on cavity convolution network |
| CN110960262A (en) * | 2019-12-31 | 2020-04-07 | 上海杏脉信息科技有限公司 | Ultrasonic scanning system, method and medium |
| CN110960262B (en) * | 2019-12-31 | 2022-06-24 | 上海杏脉信息科技有限公司 | Ultrasonic scanning system, method and medium |
| CN113438451A (en) * | 2021-06-21 | 2021-09-24 | 易成功(厦门)信息科技有限公司 | Unified standardization processing platform and method for multi-terminal multi-source data |
| CN113438451B (en) * | 2021-06-21 | 2022-04-19 | 易成功(厦门)信息科技有限公司 | Unified standardization processing platform and method for multi-terminal multi-source data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108056789A (en) | A kind of method and apparatus for the configuration parameter value for generating ultrasound scanning device | |
| US9918701B2 (en) | Methods and systems for automatic control of subjective image quality in imaging of objects | |
| US20160058426A1 (en) | Methods and systems for automatic control of subjective image quality in imaging of objects | |
| CN103648401B (en) | For evaluating the ultrasonic equipment of the quality of the osseous tissue of patient | |
| CN105338908B (en) | Ultrasonic wave optimization method and the ultrasonic therapy device for the ultrasonic wave optimization method | |
| EP2599440B1 (en) | Ultrasonic observation device, method for operating ultrasonic observation device, and operation program for ultrasonic observation device | |
| US9814447B2 (en) | Ultrasonic diagnostic apparatus | |
| CN108230261A (en) | Full-automatic image optimization based on automated organ identification | |
| CN1768709A (en) | Ultrasonic doppler measuring apparatus and control method therefor | |
| CN105246415B (en) | The method of operating of ultrasound observation apparatus and ultrasound observation apparatus | |
| CN101897600A (en) | The system and method that is used for automatic ultrasound image optimization | |
| CN103169500A (en) | Ultrasonic diagnostic apparatus, medical image diagnostic apparatus, and medical image processing method | |
| US20110077524A1 (en) | Ultrasonic diagnostic apparatus and ultrasonic contrast imaging method | |
| US8663114B2 (en) | Ultrasonic diagnostic apparatus and storage medium | |
| US11308609B2 (en) | System and methods for sequential scan parameter selection | |
| CN108652660A (en) | Diffraction correction for the decay behavior in medical diagnostic ultrasound | |
| FR3071148A1 (en) | METHODS AND SYSTEMS FOR CORRECTING SINGLE-DIMENSIONAL SHEAR WAVE DATA | |
| JP6008581B2 (en) | Ultrasonic diagnostic apparatus, control method of ultrasonic diagnostic apparatus, and ultrasonic diagnostic program | |
| CN104703543A (en) | Ultrasonic diagnosis device, sonic velocity determination method, and program | |
| CN116600697A (en) | Reflective ultrasound imaging using full waveform inversion | |
| KR20220072728A (en) | Method and system for compensating depth-dependent attenuation in ultrasonic signal data | |
| US20060184028A1 (en) | Apparatus and method for ultrasonic color imaging | |
| JP6930668B2 (en) | Ultrasonic diagnostic system | |
| CN106725607A (en) | Ultrasonic doppler parameter optimization method and ultrasonic doppler device | |
| CN104546005B (en) | A kind of ultrasound non-linear imaging method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180522 |
|
| WD01 | Invention patent application deemed withdrawn after publication |