CN117030093B - Gastric inversion support radial force measuring system for simulated press grasping machine - Google Patents
Gastric inversion support radial force measuring system for simulated press grasping machine Download PDFInfo
- Publication number
- CN117030093B CN117030093B CN202311067140.6A CN202311067140A CN117030093B CN 117030093 B CN117030093 B CN 117030093B CN 202311067140 A CN202311067140 A CN 202311067140A CN 117030093 B CN117030093 B CN 117030093B
- Authority
- CN
- China
- Prior art keywords
- radial force
- scale
- training
- vector
- neural network
- 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
- 230000002496 gastric effect Effects 0.000 title claims abstract description 93
- 238000005259 measurement Methods 0.000 claims abstract description 48
- 238000003062 neural network model Methods 0.000 claims abstract description 25
- 238000002788 crimping Methods 0.000 claims abstract description 23
- 239000013598 vector Substances 0.000 claims description 142
- 238000012549 training Methods 0.000 claims description 62
- 238000013527 convolutional neural network Methods 0.000 claims description 47
- 239000011159 matrix material Substances 0.000 claims description 41
- 238000000034 method Methods 0.000 claims description 27
- 238000000605 extraction Methods 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 17
- 230000007704 transition Effects 0.000 claims description 12
- 230000004927 fusion Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 238000012546 transfer Methods 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000011176 pooling Methods 0.000 claims description 4
- 230000000644 propagated effect Effects 0.000 claims description 2
- 230000004913 activation Effects 0.000 claims 2
- 238000005457 optimization Methods 0.000 claims 1
- 238000009826 distribution Methods 0.000 abstract description 8
- 239000000284 extract Substances 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 11
- 230000008901 benefit Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000001994 activation Methods 0.000 description 4
- 238000011068 loading method Methods 0.000 description 4
- 239000000243 solution Substances 0.000 description 4
- 210000001035 gastrointestinal tract Anatomy 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 238000003825 pressing Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 208000008589 Obesity Diseases 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 235000020824 obesity Nutrition 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01L—MEASURING FORCE, STRESS, TORQUE, WORK, MECHANICAL POWER, MECHANICAL EFFICIENCY, OR FLUID PRESSURE
- G01L5/00—Apparatus for, or methods of, measuring force, work, mechanical power, or torque, specially adapted for specific purposes
- G01L5/16—Apparatus for, or methods of, measuring force, work, mechanical power, or torque, specially adapted for specific purposes for measuring several components of force
-
- 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
- G06N3/0455—Auto-encoder networks; Encoder-decoder 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06N3/048—Activation functions
-
- 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
Abstract
A gastric inversion stent radial force measurement system simulating a crimping machine is disclosed, which extracts multi-scale characteristic information of radial force value sequence distribution of a gastric inversion stent to be measured by using a first convolution neural network model with a one-dimensional convolution kernel of a first scale and using a second convolution neural network model with a one-dimensional convolution kernel of a second scale, and generates stent radial force of the gastric inversion stent to be measured based on the multi-scale characteristic information. Thus, the accuracy of the radial force measurement of the gastric bypass stent can be improved.
Description
Technical Field
The application relates to the technical field of intelligent measurement, and more particularly relates to a gastric diversion support radial force measurement system of a simulated press-holding machine.
Background
The gastric diversion stent is a medical instrument for treating clinical obesity, and the principle is as follows: since most of the nutrition is absorbed by the intestinal tract, the gastric diversion stent is covered on a part of the intestinal tract by a film with good biocompatibility, so that the food is isolated from a part of the intestinal tract, and the absorption of the ingested food is reduced.
In the production and preparation of the gastric bypass stent, the radial force of the stent is an important detection index. Although the stent pressing machine can accurately measure the stent radial force of the gastric bypass stent, the stent pressing machine has the structure that a precise multi-piece mechanical structure simulates the cylindrical constraint of a variable diameter and records the loading force value so as to measure the stent radial force. But the stand press is an expensive piece of imported equipment and the measurement cost is high.
Thus, an optimized stent radial force measurement scheme for a gastric bypass stent is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a gastric inversion bracket radial force measurement system for simulating a crimping machine, which extracts multi-scale characteristic information of radial force value sequence distribution of a gastric inversion bracket to be measured by using a first convolution neural network model with a one-dimensional convolution kernel of a first scale and using a second convolution neural network model with a one-dimensional convolution kernel of a second scale, and generates bracket radial force of the gastric inversion bracket to be measured based on the multi-scale characteristic information. Thus, the accuracy of the radial force measurement of the gastric bypass stent can be improved.
According to one aspect of the present application, there is provided a gastric inversion support radial force measurement system simulating a crimping machine, comprising: the radial force measurement data receiving module is used for acquiring a radial force value sequence of the gastric bypass stent to be measured; the structuring module is used for arranging the radial force value sequence of the gastric diversion bracket to be measured into a radial force input vector; the first scale-associated feature extraction module is used for obtaining a first scale radial force feature vector by using a first convolution neural network model with a one-dimensional convolution kernel of a first scale; the second scale associated feature extraction module is used for obtaining a second scale radial force feature vector by using a second convolution neural network model with a one-dimensional convolution kernel of a second scale; the multi-scale fusion module is used for fusing the first-scale radial force feature vector and the second-scale radial force feature vector to obtain a decoding feature vector; and a decoding regression module for passing the decoded feature vector through a decoder to obtain a decoded value representing a stent radial force of the gastric bypass stent to be measured.
In the above-mentioned gastric diversion bracket radial force measurement system of the simulated press-holding machine, the method is characterized in that the first scale-associated feature extraction module is configured to: each layer of the first convolutional neural network model using the one-dimensional convolutional kernel having the first scale performs a one-dimensional convolutional kernel-based convolutional process, a feature matrix-based averaging process, and an activation process on input data in forward transfer of the layers, respectively, to take an output of a last layer of the first convolutional neural network model using the one-dimensional convolutional kernel having the first scale as the first-scale radial force feature vector, wherein an input of the first layer of the first convolutional neural network model using the one-dimensional convolutional kernel having the first scale is the radial force input vector.
In the above-mentioned gastric diversion support radial force measurement system of the simulation press-holding machine, the second scale associated feature extraction module is configured to: each layer of the second convolutional neural network model using the one-dimensional convolutional kernel having the second scale performs a one-dimensional convolutional kernel-based convolutional process, a feature matrix-based averaging process, and an activation process on input data in forward transfer of the layers, respectively, to take an output of a last layer of the second convolutional neural network model using the one-dimensional convolutional kernel having the second scale as the second-scale radial force feature vector, wherein an input of a first layer of the second convolutional neural network model using the one-dimensional convolutional kernel having the second scale is the radial force input vector.
In the above-mentioned gastric inversion support radial force measurement system of the simulation press-holding machine, the multi-scale fusion module further comprises: and the cascading unit is used for cascading the first-scale radial force characteristic vector and the second-scale radial force characteristic vector to obtain the decoding characteristic vector.
In the above-mentioned gastric inversion support radial force measurement system of the simulated press-and-hold machine, the decoding regression module is further configured to: performing a decoding regression on the decoded feature vector using the decoder in the following formula to obtain the decoded value; wherein, the formula is:，/>representing said decoded feature vector,/->Representing the decoded value->Representing a weight matrix, +.>Representing the bias vector +_>Representing a matrix multiplication.
The gastric diversion bracket radial force measurement system of the simulated crimping machine further comprises a training module for training the first convolutional neural network model using the one-dimensional convolutional kernel with a first scale, the second convolutional neural network model using the one-dimensional convolutional kernel with a second scale and the decoder.
In the above-mentioned gastric inversion support radial force measurement system of the simulated press-and-hold machine, the training module includes: the training data acquisition module is used for acquiring a training radial force value sequence of the gastric diversion bracket to be measured; the training structuring module is used for arranging the training radial force value sequence of the gastric diversion bracket to be measured into a training radial force input vector; a training first scale associated feature extraction module for passing the training radial force input vector through the first convolutional neural network model using a one-dimensional convolutional kernel having a first scale to obtain a training first scale radial force feature vector; a training second scale associated feature extraction module for passing the training radial force input vector through the second convolutional neural network model using a one-dimensional convolutional kernel having a second scale to obtain a training second scale radial force feature vector; the training multiscale fusion module is used for fusing the training first-scale radial force feature vector and the training second-scale radial force feature vector to obtain a training decoding feature vector; the loss function value acquisition module is used for passing the decoding characteristic vector through the decoder to obtain a decoding loss function value; the back propagation training module trains the first convolutional neural network model using the one-dimensional convolutional kernel with a first scale, the second convolutional neural network model using the one-dimensional convolutional kernel with a second scale and the decoder based on the decoding loss function value and propagated through the direction of gradient descent, wherein the weight matrix of the decoder is subject to external boundary constraint based on reference annotation in the process of each iteration.
In the above-mentioned gastric inversion support radial force measurement system of the simulated press and hold machine, performing external boundary constraint based on reference annotation on the weight matrix of the decoder includes: performing external boundary constraint based on reference annotation on the weight matrix of the decoder by using the following constraint formula to obtain an optimized weight matrix; wherein, the constraint formula is:wherein (1)>And->The weight matrix of last and current iteration, respectively,/->Is the decoding feature vector,/->Is the first transition vector, ">Is the second transition vector, ">Representing the transpose of the second transition vector, +.>Representing matrix multiplication +.>Representing matrix addition, ++>Representing the optimized weight matrix.
Compared with the prior art, the gastric diversion stent radial force measuring system of the simulated crimping machine extracts multi-scale characteristic information of radial force value sequence distribution of the gastric diversion stent to be measured by using a first convolution neural network model with a one-dimensional convolution kernel of a first scale and a second convolution neural network model with a one-dimensional convolution kernel of a second scale, and generates stent radial force of the gastric diversion stent to be measured based on the multi-scale characteristic information. Thus, the accuracy of the radial force measurement of the gastric bypass stent can be improved.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is an application scenario diagram of a gastric inversion support radial force measurement system of a simulated crimping machine according to an embodiment of the present application.
Fig. 2 is a block diagram of a gastric inversion support radial force measurement system of a simulated crimping machine in accordance with an embodiment of the present application.
FIG. 3 is a block diagram of the multi-scale fusion module in a gastric inversion support radial force measurement system of a simulated crimping machine in accordance with an embodiment of the present application.
Fig. 4 is a block diagram of the training module in a gastric inversion support radial force measurement system of a simulated press-and-grip machine according to an embodiment of the present application.
Fig. 5 is a flow chart of a method of measuring radial force of a gastric transition support of a simulated crimping machine in accordance with an embodiment of the present application.
Fig. 6 is a schematic diagram of a system architecture of a method for measuring radial force of a gastric transition support of a simulated crimping machine in accordance with an embodiment of the present application.
Fig. 7 is a schematic perspective view of the overall structure of a gastric diversion stent of a simulated crimping machine according to an embodiment of the present application.
Fig. 8 is a partial perspective view showing the internal structure of a movable cavity of a gastric diversion stent of a simulated press-and-hold machine according to an embodiment of the present application.
Fig. 9 is a bottom perspective view of a connection relationship between a gastric diversion bracket limit groove and a limit rod of a simulated crimping machine according to an embodiment of the present application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Fig. 1 is an application scenario diagram of a gastric inversion support radial force measurement system of a simulated crimping machine according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, a radial force value sequence (e.g., C illustrated in fig. 1) of a gastric bypass stent (e.g., M illustrated in fig. 1) to be measured is acquired through a polygonal cylinder structure tool (e.g., X illustrated in fig. 1). The obtained sequence of radial force values is then input into a server (e.g., S illustrated in fig. 1) that deploys a gastric bypass stent radial force measurement system algorithm of the simulated crimping machine, wherein the server is capable of processing the sequence of radial force values using the gastric bypass stent radial force measurement system algorithm of the simulated crimping machine to generate decoded values representative of stent radial forces of the gastric bypass stent to be measured.
Fig. 2 is a block diagram of a gastric inversion support radial force measurement system of a simulated crimping machine in accordance with an embodiment of the present application. As shown in fig. 2, a gastric inversion support radial force measurement system 100 of a simulated press-and-grip machine according to an embodiment of the present application includes: the radial force measurement data receiving module 110 is used for acquiring a radial force value sequence of the gastric bypass stent to be measured; a structuring module 120, configured to arrange the radial force value sequence of the gastric bypass stent to be measured into a radial force input vector; a first scale-dependent feature extraction module 130 for passing the radial force input vector through a first convolutional neural network model using a one-dimensional convolutional kernel having a first scale to obtain a first scale radial force feature vector; a second scale-dependent feature extraction module 140 for obtaining a second scale radial force feature vector from the radial force input vector by using a second convolutional neural network model having a one-dimensional convolutional kernel of a second scale; a multi-scale fusion module 150, configured to fuse the first-scale radial force feature vector and the second-scale radial force feature vector to obtain a decoded feature vector; and a decoding regression module 160 for passing the decoded feature vector through a decoder to obtain a decoded value representing the stent radial force of the gastric bypass stent to be measured.
Specifically, in the embodiment of the present application, the radial force measurement data receiving module 110 is configured to obtain a radial force value sequence of the gastric bypass stent to be measured. As described above, although the stent crimping machine can accurately measure the stent radial force of the gastric bypass stent, its structure simulates a variable diameter cylindrical constraint for a precision multi-piece mechanical structure and records the loading force value, thereby measuring the stent radial force. But the stand press is an expensive piece of imported equipment and the measurement cost is high. Thus, an optimized stent radial force measurement scheme for a gastric bypass stent is desired.
Accordingly, in the technical solution of the present application, the applicant of the present application uses a polygonal cylinder tooling to simulate a cylindrical constraint and leave a sector area as the force-measuring area, wherein the force-measuring area uses a slider structure with longitudinal guide grooves. In the radial force measurement process of the bracket, after the gastric diversion bracket is installed into a tool, the gastric diversion bracket is placed into a universal testing machine, a force between a radial force value of the bracket to be tested and a strength limit force value of the tool is set as a stopping force value of the testing machine, and a loading test is carried out on the bracket. And finally, reading a stent radial force value sequence from the loading curve, and taking the average value of the stent radial force value sequence as a measured value. And, can make the frock of different inscribed circles, can measure the radial force value that different support compression amounts correspond, through this kind of mode, reduce the measuring cost of support radial force.
In the implementation of the scheme, although the mean value is taken as the radial force measurement value of the gastric bypass stent and can reflect the true radial force of the gastric bypass stent to a certain extent, the mean value is basically an approximate estimated value of the radial force of the stent, and the radial force of the stent cannot be accurately characterized. And the consideration of the stent radial force value sequence can reflect the performance and the characteristics of the stent, so that if the characteristic information implicit in the stent radial force value sequence can be more fully utilized, the measurement accuracy of the stent radial force can be improved.
Specifically, a sequence of radial force values of the gastric bypass stent to be measured is first acquired. It should be appreciated that the sequence of radial force values of the gastric bypass stent to be measured may include a plurality of radial force values of the gastric bypass stent to be measured so that the final decoded value is more accurate.
Specifically, in the embodiment of the present application, the structuring module 120 is configured to arrange the radial force value sequence of the gastric bypass stent to be measured into a radial force input vector. After the radial force value sequence of the gastric bypass stent to be measured is obtained, the radial force value sequence of the gastric bypass stent to be measured is arranged as a radial force input vector. That is, the radial force value sequence is vectorized at the data structure level to obtain the radial force input vector.
In this way, the radial force input vector can be input into a computer, which is beneficial to the subsequent processing of data in a convolutional neural network.
Specifically, in the embodiment of the present application, the first scale-associated feature extraction module 130 is configured to obtain the first scale radial force feature vector by using a first convolutional neural network model with a one-dimensional convolutional kernel of a first scale. The radial force input vector is then passed through a first convolutional neural network model using a one-dimensional convolutional kernel having a first scale to obtain a first scale radial force feature vector. That is, a convolutional neural network model with excellent performance in the local feature extraction field is used as a local correlation feature extractor to capture high-dimensional local implicit features in the radial force input vector.
More specifically, in the technical solution of the present application, the first convolutional neural network model is a one-dimensional convolutional neural network model, which performs one-dimensional convolutional encoding on the radial force input vector by using a one-dimensional convolutional layer with a first scale to extract high-dimensional correlation implicit features between a plurality of radial force measurement values of the radial force input vector within different sample spans.
Further, in the embodiment of the application, each layer of the first convolutional neural network model using the one-dimensional convolutional kernel with the first scale performs a convolutional process based on the one-dimensional convolutional kernel, a mean pooling process based on the feature matrix and an activation process on input data in forward transfer of the layer respectively to take output of the last layer of the first convolutional neural network model using the one-dimensional convolutional kernel with the first scale as the first-scale radial force feature vector, wherein input of the first layer of the first convolutional neural network model using the one-dimensional convolutional kernel with the first scale is the radial force input vector.
It should be appreciated that the high-dimensional local implicit features in the radial force input vector are captured by using a convolutional neural network model with excellent performance in the field of local feature extraction as a local associative feature extractor to obtain a first-scale radial force feature vector.
Specifically, in the embodiment of the present application, the second scale-associated feature extraction module 140 is configured to obtain the second scale radial force feature vector by using a second convolutional neural network model with a one-dimensional convolutional kernel of a second scale. The first convolutional neural network model can only capture the associated information of the predetermined scale of the radial force input vector, which is not felt by the associated information outside the predetermined scale, limited by the scale of the one-dimensional convolutional kernel of the first convolutional neural network model.
In order to expand the feature receptive field, in the technical solution of the application, the radial force input vector is further obtained through a second convolution neural network model with a one-dimensional convolution kernel of a second scale, so as to obtain a radial force feature vector of the second scale.
Here, the scale of the one-dimensional convolution kernel used by the second convolution neural network model is different from the scale of the one-dimensional convolution kernel used by the first convolution neural network model, so that the second convolution neural network model and the second convolution neural network model have different feature receptive fields, that is, multi-scale feature information of the sequence distribution of the radial force input vector can be extracted by comprehensively utilizing the first convolution neural network model and the second convolution neural network model.
Further, in the embodiment of the application, each layer of the second convolutional neural network model using the one-dimensional convolutional kernel with the second scale performs a convolutional process based on the one-dimensional convolutional kernel, a mean pooling process based on the feature matrix and an activation process on the input data in forward transfer of the layer respectively to take the output of the last layer of the second convolutional neural network model using the one-dimensional convolutional kernel with the second scale as the second-scale radial force feature vector, wherein the input of the first layer of the second convolutional neural network model using the one-dimensional convolutional kernel with the second scale is the radial force input vector.
It should be appreciated that by capturing high-dimensional local implicit features in the radial force input vector using a convolutional neural network model with excellent performance in the local feature extraction field as a local correlation feature extractor, multi-scale feature information of the sequence distribution of the radial force input vector can be extracted by comprehensively utilizing the first convolutional neural network model and the second convolutional neural network model, so as to obtain a first-scale radial force feature vector and a second-scale radial force feature vector.
Specifically, in the embodiment of the present application, the multi-scale fusion module 150 is configured to fuse the first-scale radial force feature vector and the second-scale radial force feature vector to obtain a decoded feature vector. And then, after the first-scale radial force characteristic vector and the second-scale radial force characteristic vector are fused to obtain a decoding characteristic vector, the decoding characteristic vector is passed through a decoder to obtain a decoding value which has higher measurement accuracy and is used for representing the stent radial force of the gastric bypass stent to be measured.
Specifically, in the embodiment of the present application, the decoding regression module 160 is configured to pass the decoding eigenvector through a decoder to obtain a decoded value for representing the stent radial force of the gastric bypass stent to be measured. Thus, the accuracy of the radial force measurement of the gastric bypass stent can be improved.
Further, performing decoding regression on the decoded feature vector using the decoder in the following formula to obtain the decoded value; wherein, the formula is:，/>the decoded feature vector is represented as such,representing the decoded value->Representing a weight matrix, +.>Representing the bias vector +_>Representing a matrix multiplication.
Further, fig. 4 is a block diagram of the training module in the gastric inversion support radial force measurement system of the simulated press and hold machine according to an embodiment of the present application, and as shown in fig. 4, the gastric inversion support radial force measurement system 100 of the simulated press and hold machine further includes a training module 300 for training the first convolutional neural network model using the one-dimensional convolutional kernel having the first scale, the second convolutional neural network model using the one-dimensional convolutional kernel having the second scale, and the decoder. Wherein the training module 300 comprises: a training data acquisition module 310, configured to acquire a training radial force value sequence of the gastric bypass stent to be measured; a training structuring module 320, configured to arrange the training radial force value sequence of the gastric bypass stent to be measured into a training radial force input vector; a training first scale associated feature extraction module 330 for passing the training radial force input vector through the first convolutional neural network model using a one-dimensional convolutional kernel having a first scale to obtain a training first scale radial force feature vector; a training second scale associated feature extraction module 340 for passing the training radial force input vector through the second convolutional neural network model using a one-dimensional convolutional kernel having a second scale to obtain a training second scale radial force feature vector; a training multiscale fusion module 350, configured to fuse the training first-scale radial force feature vector and the training second-scale radial force feature vector to obtain a training decoding feature vector; a loss function value obtaining module 360, configured to pass the decoded feature vector through the decoder to obtain a decoding loss function value; the back propagation training module 370 trains the first convolutional neural network model using the one-dimensional convolutional kernel having a first scale, the second convolutional neural network model using the one-dimensional convolutional kernel having a second scale, and the decoder based on the decoding loss function value and propagation through a direction of gradient descent, wherein, during each iteration, an external boundary constraint based on a reference annotation is applied to a weight matrix of the decoder.
In the technical solution of the present application, since the first-scale radial force feature vector and the second-scale radial force feature vector respectively represent associated distribution features of radial force values in different scales along a time sequence dimension, in order to fully utilize the above information to decode, the decoded feature vector is preferably obtained by concatenating the first-scale radial force feature vector and the second-scale radial force feature vector. However, the first-scale radial force feature vector and the second-scale radial force feature vector are respectively used as local feature distribution, so that the time sequence dimension expression of the decoding feature vector deviates from the source time sequence feature expression of a radial force value under the direct cascade connection condition, thereby causing time sequence feature domain deviation of regression probability mapping of the decoding feature vector in the weight matrix iteration process of a decoder under the decoding scene, and further causing the weight matrix to diverge based on time sequence fitting of the decoding feature vector, thereby influencing the training effect of a model and the accuracy of the decoding result of the decoding feature vector obtained by the trained model.
Based on the above, the applicant of the present application performs reference-based injection of weight matrix during the training process of the decoded feature vector through the decoderThe released external boundary constraint is specifically expressed as:wherein (1)>And->The weight matrix of the last iteration and the current iteration are respectively adopted, wherein, during the first iteration, different initialization strategies are adopted to set +.>And->(e.g.)>Set as a unitary matrix->Set as the diagonal matrix of the mean value of the feature vector to be classified),>is the said decoded feature vector(s), in the form of column vectors>Is the first transition vector, ">Is the second transition vector, ">Representing the transpose of the second transition vector, +.>Representing matrix multiplication +.>Representing matrix addition, ++>Representing the optimized weight matrix.
Here, by decoding the feature vector as describedThe iterative association expression in the weight space is used as the external association boundary constraint of the weight matrix iteration, so that the decoding eigenvector +_is reduced in the weight space iteration process under the condition that the previous weight matrix is used as the reference annotation (benchmark annotation) in the current iteration process>Is used as an anchor point, thereby performing a weight matrix in an iterative process relative to the decoding eigenvector ++ oriented mismatch>Compensation of timing offset of the regression probability map of (2), and further enhancing the weight matrix based on the decoded feature vector +.>To promote the training effect of the model and the accuracy of the decoding result of the decoding feature vector obtained by the trained model.
In summary, a gastric inversion stent radial force measurement system 100 of a simulated crimping machine in accordance with an embodiment of the present application is illustrated that extracts multi-scale feature information of a radial force value sequence distribution of a gastric inversion stent to be measured by using a first convolutional neural network model having one-dimensional convolutional kernels of a first scale and using a second convolutional neural network model having one-dimensional convolutional kernels of a second scale, and generates stent radial forces of the gastric inversion stent to be measured based thereon. Thus, the accuracy of the radial force measurement of the gastric bypass stent can be improved.
Exemplary method fig. 5 is a flow chart of a method of measuring radial force of a gastric bypass stent simulating a crimping machine in accordance with an embodiment of the present application. Fig. 6 is a schematic diagram of a system architecture of a method for measuring radial force of a gastric transition support of a simulated crimping machine in accordance with an embodiment of the present application. As shown in fig. 5 and 6, a method for measuring radial force of a gastric bypass stent of a simulated press-and-grip machine according to an embodiment of the present application includes: s110, acquiring a radial force value sequence of a gastric shunt support to be measured; s120, arranging the radial force value sequence of the gastric bypass stent to be measured into a radial force input vector; s130, the radial force input vector is obtained through a first convolution neural network model with a one-dimensional convolution kernel of a first scale, so as to obtain a radial force characteristic vector of the first scale; s140, the radial force input vector is obtained through a second convolution neural network model with a one-dimensional convolution kernel of a second scale, so as to obtain a radial force characteristic vector of the second scale; s150, fusing the first-scale radial force feature vector and the second-scale radial force feature vector to obtain a decoding feature vector; and S160, passing the decoding characteristic vector through a decoder to obtain a decoding value for representing the stent radial force of the gastric bypass stent to be measured.
Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described method for measuring the radial force of the gastric bypass stent of the simulated crimping machine have been described in detail in the gastric bypass stent radial force measuring system of the simulated crimping machine with reference to fig. 1 to 4, and thus, repetitive descriptions thereof will be omitted.
Structural example this example relates to the following structure: the problem that the deformation and cracking cannot be caused due to overlarge force in the measuring process is solved by the aid of the polygonal column tool X, the workbench X1, the fixing seat X2, the movable cavity X3, the guide rod X4, the spring X5, the tray X6, the limiting groove X7, the fixing shaft X8, the protection plate X9 and the limiting rod X10.
Referring to fig. 7 to 9, it is shown that: the specific working process is as follows: during operation, place polygon column frock X on tray X6, at this moment under the effect of top dynamometer side head pushing down, polygon column frock X pushes down tray X6 when receiving the power of pushing down, tray X6's both sides down move and compress spring X5 on guide arm X4, tray X6 passes through limit groove X7 and drives gag lever post X10's inboard end down movement, gag lever post X10 drives protection shield X9 and uses fixed axle X8 to rotate as the axle centre downward inboard, and then realize the guard action to polygon column frock X, can prevent to have the foreign object to influence measurement work at the in-process of measurement, also prevent simultaneously that measuring equipment from breaking apart polygon column frock X too big and leading to polygon column frock X's problem when breaking apart, prevent that polygon column frock X from causing the striking to surrounding equipment, the security of equipment at the during operation improves.
The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.
The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.
It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (4)
1. A gastric inversion support radial force measurement system for a simulated crimping machine, comprising: the radial force measurement data receiving module is used for acquiring a radial force value sequence of the gastric bypass stent to be measured; the structuring module is used for arranging the radial force value sequence of the gastric diversion bracket to be measured into a radial force input vector; the first scale-associated feature extraction module is used for obtaining a first scale radial force feature vector by using a first convolution neural network model with a one-dimensional convolution kernel of a first scale; the second scale associated feature extraction module is used for obtaining a second scale radial force feature vector by using a second convolution neural network model with a one-dimensional convolution kernel of a second scale; the multi-scale fusion module is used for fusing the first-scale radial force feature vector and the second-scale radial force feature vector to obtain a decoding feature vector; the decoding regression module is used for passing the decoding characteristic vector through a decoder to obtain a decoding value used for representing the stent radial force of the gastric bypass stent to be measured;
the first scale-associated feature extraction module is used for: performing one-dimensional convolution kernel-based convolution processing, feature matrix-based averaging pooling processing and activation processing on input data in forward transfer of layers respectively by using layers of a first convolution neural network model with a one-dimensional convolution kernel of a first scale to output a first-scale radial force feature vector by the last layer of the first convolution neural network model with the one-dimensional convolution kernel of the first scale, wherein input of the first layer of the first convolution neural network model with the one-dimensional convolution kernel of the first scale is the radial force input vector;
the second scale associated feature extraction module is configured to: performing one-dimensional convolution kernel-based convolution processing, feature matrix-based averaging pooling processing and activation processing on input data in forward transfer of layers respectively by using layers of a second convolution neural network model with a one-dimensional convolution kernel of a second scale to take output of a last layer of the second convolution neural network model with the one-dimensional convolution kernel of the second scale as the second-scale radial force feature vector, wherein input of a first layer of the second convolution neural network model with the one-dimensional convolution kernel of the second scale is the radial force input vector;
wherein, multiscale fuses the module, still includes: the cascade unit is used for cascading the first-scale radial force characteristic vector and the second-scale radial force characteristic vector to obtain the decoding characteristic vector;
wherein, the decoding regression module is used for: performing a decoding regression on the decoded feature vector using the decoder in the following formula to obtain the decoded value; wherein, the formula is: V d representing the decoding eigenvector, Y representing the decoded value, W representing the weight matrix, B representing the bias vector, +.>Representing a matrix multiplication.
2. The gastric inversion stent radial force measurement system of claim 1 further comprising a training module for training the first convolutional neural network model using a one-dimensional convolutional kernel having a first scale, the second convolutional neural network model using a one-dimensional convolutional kernel having a second scale, and the decoder.
3. The gastric bypass stent radial force measurement system of a simulated press and hold machine of claim 2, wherein the training module comprises: the training data acquisition module is used for acquiring a training radial force value sequence of the gastric diversion bracket to be measured; the training structuring module is used for arranging the training radial force value sequence of the gastric diversion bracket to be measured into a training radial force input vector; a training first scale associated feature extraction module for passing the training radial force input vector through the first convolutional neural network model using a one-dimensional convolutional kernel having a first scale to obtain a training first scale radial force feature vector; a training second scale associated feature extraction module for passing the training radial force input vector through the second convolutional neural network model using a one-dimensional convolutional kernel having a second scale to obtain a training second scale radial force feature vector; the training multiscale fusion module is used for fusing the training first-scale radial force feature vector and the training second-scale radial force feature vector to obtain a training decoding feature vector; the loss function value acquisition module is used for passing the decoding characteristic vector through the decoder to obtain a decoding loss function value; the back propagation training module trains the first convolutional neural network model using the one-dimensional convolutional kernel with a first scale, the second convolutional neural network model using the one-dimensional convolutional kernel with a second scale and the decoder based on the decoding loss function value and propagated through the direction of gradient descent, wherein the weight matrix of the decoder is subject to external boundary constraint based on reference annotation in the process of each iteration.
4. A gastric bypass stent radial force measurement system of a simulated press and hold machine as claimed in claim 3 wherein reference annotation based external boundary constraints are applied to the weight matrix of the decoder comprising: performing external boundary constraint based on reference annotation on the weight matrix of the decoder by using the following constraint formula to obtain an optimized weight matrix; wherein, the constraint formula is:
wherein M is 1 And M 2 The weight matrix of the last iteration and the current iteration are respectively V c Is the decoded feature vector, V 1 Is the first transition vector, V 2 Is the second transition vector, V 2 T Representing a transpose of the second transition vector,representing matrix multiplication +.>Representing matrix addition, M 2 ' represents the optimization weight matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311067140.6A CN117030093B (en) | 2023-08-23 | 2023-08-23 | Gastric inversion support radial force measuring system for simulated press grasping machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311067140.6A CN117030093B (en) | 2023-08-23 | 2023-08-23 | Gastric inversion support radial force measuring system for simulated press grasping machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117030093A CN117030093A (en) | 2023-11-10 |
| CN117030093B true CN117030093B (en) | 2024-01-26 |
Family
ID=88637170
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311067140.6A Active CN117030093B (en) | 2023-08-23 | 2023-08-23 | Gastric inversion support radial force measuring system for simulated press grasping machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117030093B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117272839B (en) * | 2023-11-20 | 2024-02-06 | 北京阿迈特医疗器械有限公司 | Support press-holding performance prediction method and device based on neural network |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1856699A (en) * | 2003-09-24 | 2006-11-01 | Skf股份公司 | Method and sensor arrangement for load measurement on rolling element bearing based on model deformation |
| CN113267288A (en) * | 2021-06-08 | 2021-08-17 | 上海理工大学 | Radial supporting force test experiment platform and method for thoracic aorta covered stent |
| CN216770866U (en) * | 2021-10-15 | 2022-06-17 | 上海盈阁智能科技有限公司 | Press-holding mechanism for radial force test |
| WO2022136836A1 (en) * | 2020-12-21 | 2022-06-30 | Ucl Business Ltd | A system for positioning a medical tool |
| CN115831373A (en) * | 2022-10-28 | 2023-03-21 | 广西大学 | Simulation method for polymer esophageal stent implantation |
| CN115887086A (en) * | 2023-01-10 | 2023-04-04 | 杭州糖吉医疗科技有限公司 | Controllable steering gastric bypass stent delivery and method thereof |
-
2023
- 2023-08-23 CN CN202311067140.6A patent/CN117030093B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1856699A (en) * | 2003-09-24 | 2006-11-01 | Skf股份公司 | Method and sensor arrangement for load measurement on rolling element bearing based on model deformation |
| WO2022136836A1 (en) * | 2020-12-21 | 2022-06-30 | Ucl Business Ltd | A system for positioning a medical tool |
| CN113267288A (en) * | 2021-06-08 | 2021-08-17 | 上海理工大学 | Radial supporting force test experiment platform and method for thoracic aorta covered stent |
| CN216770866U (en) * | 2021-10-15 | 2022-06-17 | 上海盈阁智能科技有限公司 | Press-holding mechanism for radial force test |
| CN115831373A (en) * | 2022-10-28 | 2023-03-21 | 广西大学 | Simulation method for polymer esophageal stent implantation |
| CN115887086A (en) * | 2023-01-10 | 2023-04-04 | 杭州糖吉医疗科技有限公司 | Controllable steering gastric bypass stent delivery and method thereof |
Non-Patent Citations (1)
| Title |
|---|
| 基于非对称孔结构的高性能血管支架设计;周文选;王明;高旻昱;李祥;罗云;;医用生物力学(06);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117030093A (en) | 2023-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117030093B (en) | Gastric inversion support radial force measuring system for simulated press grasping machine | |
| CN113610151B (en) | Small sample image classification system based on prototype network and self-encoder | |
| CN110148103B (en) | Hyperspectral and multispectral image fusion method based on joint optimization, computer-readable storage medium and electronic device | |
| CN109993105A (en) | A kind of improved adaptive sparse sampling Fault Classification | |
| Chen et al. | Compressive sensing for fast analysis of wide-angle monostatic scattering problems | |
| WO2019153566A1 (en) | Method and device for magnetic resonance parameter imaging, medical equipment, and storage medium | |
| CN111951344A (en) | Magnetic resonance image reconstruction method based on cascade parallel convolution network | |
| CN107862170B (en) | Finite element model correction method based on dynamic polycondensation | |
| CN109886337B (en) | Depth measurement learning method and system based on self-adaptive sampling | |
| CN106501865A (en) | A kind of sparse imaging method of the nested weighting in edge | |
| CN107092805B (en) | Magnetic resonance parallel imaging device | |
| CN111595541A (en) | Multi-dimensional structure damage identification method based on mass vibration transmissibility data convolutional neural network processing | |
| Häußler et al. | Galapagos: A Semi-Automated Tool for Galaxy Profile Fitting | |
| Malis et al. | Compressed sensing velocity encoded phase contrast imaging: monitoring skeletal muscle kinematics | |
| CN113109780B (en) | High-resolution range profile target identification method based on complex number dense connection neural network | |
| CN117233682B (en) | Quick calibration system of balance bridge | |
| CN116612335B (en) | Few-sample fine-granularity image classification method based on contrast learning | |
| CN117058149A (en) | Method for training and identifying medical image measurement model of osteoarthritis | |
| CN105759233B (en) | A kind of rapid chemical exchanges saturation transfer imaging method and system | |
| CN117095236A (en) | Method and system for evaluating test accuracy of blade root wheel groove | |
| CN116858789A (en) | Food safety detection system and method thereof | |
| CN111798531A (en) | Image depth convolution compressed sensing reconstruction method applied to plant monitoring | |
| CN116124448A (en) | Fault diagnosis system and method for wind power gear box | |
| Cetindag et al. | Transfer Learning Methods for Using Textural Features in Histopathological Image Classification | |
| CN115641263A (en) | Single-power equipment infrared image super-resolution reconstruction method based on deep learning |
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 |