CN113487581A - Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks - Google Patents
Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks Download PDFInfo
- Publication number
- CN113487581A CN113487581A CN202110806990.8A CN202110806990A CN113487581A CN 113487581 A CN113487581 A CN 113487581A CN 202110806990 A CN202110806990 A CN 202110806990A CN 113487581 A CN113487581 A CN 113487581A
- Authority
- CN
- China
- Prior art keywords
- image
- fetus
- fetal
- ultrasonic
- training
- 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
- 210000003754 fetus Anatomy 0.000 title claims abstract description 149
- 238000000034 method Methods 0.000 title claims abstract description 70
- 210000001217 buttock Anatomy 0.000 title description 2
- 230000001605 fetal effect Effects 0.000 claims abstract description 117
- 238000003709 image segmentation Methods 0.000 claims abstract description 97
- 238000012549 training Methods 0.000 claims abstract description 83
- 238000002604 ultrasonography Methods 0.000 claims description 35
- 230000008569 process Effects 0.000 claims description 32
- 238000012545 processing Methods 0.000 claims description 31
- 238000005070 sampling Methods 0.000 claims description 27
- 230000006870 function Effects 0.000 claims description 24
- 238000005259 measurement Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 9
- 238000000691 measurement method Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 238000010008 shearing Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 3
- 230000001413 cellular effect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000035935 pregnancy Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 210000001015 abdomen Anatomy 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 210000003484 anatomy Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 210000000689 upper leg Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30044—Fetus; Embryo
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The application relates to a method, a system, equipment and a storage medium for automatically measuring the head and hip diameter of a fetus, wherein the method comprises the steps of acquiring an ultrasonic training image of the fetus and a corresponding label image, wherein the interested area of the label image is the whole body of the fetus; inputting the fetal ultrasonic training image and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model; acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to the image segmentation target model to obtain a predicted fetus image; and calculating the length of the hip diameter of the fetus head according to the predicted fetus image. The method and the device are helpful for quickly and accurately calculating the length of the hip diameter of the fetal head.
Description
Technical Field
The application relates to the technical field of ultrasonic image processing, in particular to a method, a system, equipment and a storage medium for automatically measuring the hip diameter of a fetal head.
Background
When a newborn is born, if physical defects can cause serious diseases and even death of the fetus, heavy attacks and burdens are brought to the family of the newborn. The prenatal examination can find the fetal defects as soon as possible, and has higher application value for judging whether to continue pregnancy. Through ultrasonic examination, not only can some defects of the fetus be found in time, but also some anatomical structures of the fetus can be measured, and the method has very important significance for predicting the pregnancy of the pregnant woman and estimating the size and weight of the fetus.
Obstetrical ultrasonic examination mainly measures parameters such as head circumference, abdomen circumference, femur length, head-hip diameter and the like of a fetus based on an ultrasonic image, and the measurement of the biological parameters can be used for judging whether the growth condition of the fetus is good and whether the fetus is abnormal. Now, these parameters need to be measured manually by a doctor or automatically by a conventional image processing algorithm, the former is inefficient and depends heavily on the doctor's experience, the latter algorithm is not easy to implement, and the accuracy rate needs to be improved, so the inventor thinks that there is still further improvement in fetal ultrasound image processing.
Disclosure of Invention
In view of this, the present application provides a method, a system, a device and a storage medium for automatically measuring a diameter of a head and a hip of a fetus, so as to solve the technical problems of low efficiency and low accuracy of the existing detection of the diameter of the head and the hip of the fetus.
In order to solve the above problem, in a first aspect, the present application provides a method for automatically measuring a hip diameter of a fetus, the method comprising:
acquiring a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is the whole body of the fetus;
inputting the fetal ultrasonic training image and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to the image segmentation target model to obtain a predicted fetus image;
and calculating the length of the hip diameter of the fetus head according to the predicted fetus image.
Optionally, the acquiring the ultrasound training image of the fetus and the corresponding tag image includes:
acquiring a plurality of fetal ultrasonic initial images, and shearing boundary interference areas of all the fetal ultrasonic initial images to obtain corresponding fetal ultrasonic target images;
performing data enhancement processing on all the fetal ultrasonic target images to obtain fetal ultrasonic training images;
backing up all the fetal ultrasonic target images, and marking a fetal whole body in the backed-up fetal ultrasonic target images as an interested area;
and performing data enhancement processing on all marked ultrasonic target images of the fetus to obtain corresponding label images.
Optionally, marking a whole fetus in the backed-up fetus ultrasound target image as a region of interest, wherein a gray value of the region of interest is 1; and marking the region except the whole body of the fetus in the backed-up fetus ultrasonic target image as a region of no interest, wherein the gray value of the region of no interest is 0.
Optionally, before the fetal ultrasound training image and the corresponding tag image are input to a preset image segmentation initial model, the method further includes:
improving a preset Unet network to obtain the image segmentation initial model, including:
replacing part of a plurality of convolution layers in the downsampling process of the Unet network with expansion convolution layers, wherein the expansion rates of the plurality of expansion convolution layers are gradually increased along with the increase of the hierarchy;
setting batch normalization for each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network;
and setting each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network as residual network connection.
Optionally, the improving a preset Unet network to obtain the image segmentation initial model further includes:
setting a Leaky-relu activation function for each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network;
and adding a sigmoid activation function and a sigmoid cross entropy loss function to the last layer of the Unet network.
Optionally, the inputting the fetal ultrasound training image and the corresponding tag image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model, includes:
inputting the fetal ultrasonic training image into an image segmentation initial model for training, outputting a predicted image, and calculating a loss value between the predicted image and a label image corresponding to the fetal ultrasonic training image;
and based on the loss value, adopting a gradient descent algorithm to carry out iterative training, minimizing a loss function, updating network parameters of the image segmentation initial model, wherein the network parameters comprise the weight and the offset of each layer of network, and taking the image segmentation initial model after the convergence training as an image segmentation target model.
Optionally, calculating the head-hip diameter length of the fetus according to the predicted fetus image includes:
performing binarization processing on the predicted fetal image, wherein the corresponding processing of pixel values of the predicted fetal image, which are greater than a preset threshold value, into 1 and the corresponding processing of pixel values of the predicted fetal image, which are less than or equal to the preset threshold value, into 0;
and fitting the predicted fetal image boundary after the binarization processing, determining two farthest end points in the points with the pixel value of 1, and taking the distance between the two farthest end points as the length of the hip diameter of the fetal head.
In a second aspect, the present application further provides a fetal head-hip diameter automatic measurement system, the system comprising:
the image acquisition module is used for acquiring a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is a whole fetus body;
the training module is used for inputting the fetal ultrasonic training image and the corresponding label image into a preset image segmentation initial model and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
the prediction module is used for acquiring an ultrasonic image of the fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to the image segmentation target model to obtain a predicted fetus image;
and the calculation module is used for calculating the length of the head and the hip diameter of the fetus according to the predicted fetus image.
In a third aspect, the present application provides a computer device, which adopts the following technical solution:
a computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method for automatic measurement of the fetal head-hip diameter when executing the computer program.
In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solutions:
a computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method for automatic measurement of the diameter of the head and the hip of a fetus.
The beneficial effects of adopting the above embodiment are: obtaining a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is a whole fetus body, so that the whole fetus body can be conveniently trained and segmented by a subsequent image segmentation initial model; carrying out iterative convergence training on the image segmentation initial model to obtain a well-trained image segmentation target model, thereby facilitating image segmentation on the ultrasonic training image of the fetus by calling the model medically; the ultrasonic image of the fetus to be detected is input into the image segmentation target model to obtain a predicted fetus image, and the head-hip diameter length of the fetus can be rapidly and accurately calculated according to the more accurate shape of the whole body of the fetus in the predicted fetus image.
Drawings
Fig. 1 is a schematic view of an application scenario of an automatic fetal head-hip diameter measurement system provided in the present application;
fig. 2 is a flowchart of an embodiment of an automatic fetal head-hip diameter measurement method provided by the present application;
fig. 3 is a flowchart illustrating an embodiment of a step S201 of the automatic fetal head-hip diameter measurement method provided in the present application;
fig. 4(a), fig. 4(b), and fig. 4(c) are schematic diagrams of three ultrasound training images of a fetus provided by the present application, respectively;
fig. 5(a), 5(b) and 5(c) are respectively label images corresponding to fig. 4(a), 4(b) and 4(c) which are corresponding to three fetal ultrasound training images;
fig. 6 is a flowchart of another embodiment of an automatic fetal head-hip diameter measurement method provided by the present application;
FIG. 7 is a model architecture diagram of an embodiment of an initial model for image segmentation provided herein;
FIG. 8 is a diagram illustrating parameter settings of layers in an embodiment of an initial model for image segmentation provided in the present application;
FIG. 9 is a schematic diagram of an embodiment of a residual error network architecture provided herein;
fig. 10 is a flowchart illustrating an exemplary method for automatically measuring the hip diameter of the fetal head according to the step S601;
fig. 11 is a flowchart illustrating an exemplary method for automatically measuring the hip diameter of the fetal head according to step S202;
fig. 12 is a flowchart illustrating an exemplary method for automatically measuring the hip diameter of the fetal head according to step S204;
fig. 13(a), fig. 13(b), and fig. 13(c) are schematic diagrams of the fetal ultrasound target image in the three test sets provided in the present application, respectively;
fig. 14(a), 14(b) and 14(c) correspond to the predicted fetal images predicted and binarized in fig. 13(a), 13(b) and 13(c), respectively;
fig. 15(a), 15(b) and 15(c) are schematic diagrams corresponding to fig. 13(a), 13(b) and 13(c) respectively for measuring the diameter of the head and the hip of the fetus;
fig. 16 is a schematic block diagram of an embodiment of an automatic fetal head-hip diameter measurement system provided by the present application;
FIG. 17 is a functional block diagram of an embodiment of a computer device provided herein.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the application and together with the description, serve to explain the principles of the application and not to limit the scope of the application.
In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The application provides a method, a system, equipment and a storage medium for automatically measuring the diameter of the head and the hip of a fetus, which are respectively explained in detail below.
Fig. 1 is a schematic view of a scenario of an automatic fetal head-hip diameter measuring system provided in an embodiment of the present application, where the system may include a server 100, and the server 100 integrates the automatic fetal head-hip diameter measuring system, such as the server in fig. 1.
In the embodiment of the present application, the server 100 is mainly used for:
acquiring a fetus ultrasonic training image and a corresponding label image, wherein the region of interest of the label image is the whole body of the fetus;
inputting the ultrasonic training image of the fetus and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected into an image segmentation target model to obtain a predicted fetus image;
and calculating the length of the hip diameter of the head of the fetus according to the predicted fetus image.
In this embodiment, the server 100 may be an independent server, or may be a server network or a server cluster composed of servers, for example, the server 100 described in this embodiment includes, but is not limited to, a computer, a network host, a single network server, a plurality of network server sets, or a cloud server composed of a plurality of servers. Among them, the Cloud server is constituted by a large number of computers or web servers based on Cloud Computing (Cloud Computing).
It is to be understood that the terminal 200 used in the embodiments of the present application may be a device that includes both receiving and transmitting hardware, i.e., a device having receiving and transmitting hardware capable of performing two-way communication over a two-way communication link. Such a device may include: a cellular or other communication device having a single line display or a multi-line display or a cellular or other communication device without a multi-line display. The specific terminal 200 may be a desktop, a laptop, a web server, a Personal Digital Assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, an embedded device, and the like, and the type of the terminal 200 is not limited in this embodiment.
It will be understood by those skilled in the art that the application environment shown in fig. 1 is only one application scenario related to the present application, and does not constitute a limitation on the application scenario of the present application, and that other application environments may further include more or fewer terminals than those shown in fig. 1, for example, only 2 terminals are shown in fig. 1, and it is understood that the automatic fetal head-hip diameter measurement system may further include one or more other terminals, which are not limited herein.
In addition, referring to fig. 1, the automatic fetal head-hip diameter measuring system may further include a memory 200 for storing data, such as data of the ultrasonic training image of the fetus, the tag image, and the model parameters of the image segmentation target model.
It should be noted that the scenario diagram of the automatic fetal head-hip diameter measurement system shown in fig. 1 is merely an example, and the automatic fetal head-hip diameter measurement system and the scenario described in the embodiment of the present application are for more clearly illustrating the technical solution of the embodiment of the present application, and do not form a limitation on the technical solution provided in the embodiment of the present application.
Referring to fig. 2, a flowchart of a method of an embodiment of the automatic fetal head-hip diameter measurement method provided by the present application includes the following steps:
s201, obtaining a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is a whole fetus body;
s202, inputting the ultrasonic training image of the fetus and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
s203, obtaining an ultrasonic image of the fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to an image segmentation target model to obtain a predicted fetus image;
and S204, calculating the head-hip diameter length of the fetus according to the predicted fetus image.
It should be noted that the fetal ultrasound training image may be a fetal color ultrasound or B-ultrasound image.
According to the embodiment, the ultrasonic training image of the fetus and the corresponding label image are obtained, wherein the interested region of the label image is the whole body of the fetus, so that the whole body of the fetus can be conveniently trained and segmented by the initial model for segmenting the subsequent images; carrying out iterative convergence training on the image segmentation initial model to obtain a well-trained image segmentation target model, thereby facilitating image segmentation on the ultrasonic training image of the fetus by calling the model medically; the ultrasonic image of the fetus to be detected is input into the image segmentation target model to obtain a predicted fetus image, and the head-hip diameter length of the fetus can be rapidly and accurately calculated according to the more accurate shape of the whole body of the fetus in the predicted fetus image.
Referring to fig. 3, a flowchart of a method of an embodiment of step S201 provided in the present application, where the step S201 includes the following steps:
s301, acquiring a plurality of fetal ultrasonic initial images, and shearing boundary interference areas of all the fetal ultrasonic initial images to obtain corresponding fetal ultrasonic target images;
s302, performing data enhancement processing on all the fetal ultrasonic target images to obtain fetal ultrasonic training images;
s303, backing up all the fetal ultrasonic target images, and marking the whole fetal body in the backed-up fetal ultrasonic target images as an interested area;
s304, performing data enhancement processing on all marked fetal ultrasonic target images to obtain corresponding label images.
In this embodiment, the fetal ultrasound initial image may refer to a fetal B-mode ultrasound image or a color ultrasound image. The fetal ultrasound target image may refer to an image that retains a fetal ultrasound detection image area.
In a specific embodiment, 700 fetal ultrasound initial images are selected, shearing processing is performed, image areas of all the images are reserved, and boundary interference areas are removed to obtain a fetal ultrasound target image. Further, all the fetal ultrasound target images are backed up, and then each fetal ultrasound target image is labeled by a professional sonographer, and the whole body of the fetus is labeled as a region of interest (ROI) to obtain a label image.
Further, 600 fetal ultrasound target images are selected as a training set, the remaining 100 images are used as a test set, the 600 fetal ultrasound training images are expanded to 3000 images through data enhancement processing, and tag images corresponding to the 600 fetal ultrasound training images are also expanded to 3000 images through data enhancement processing. Exemplarily, three ultrasound training images of the fetus are shown in fig. 4(a), fig. 4(b) and fig. 4 (c).
It should be noted that, in other embodiments, the number of the fetal ultrasound initial images and the division ratio of the training set and the test set may be determined according to actual situations, and is not limited herein.
In an embodiment, in step S303, the whole body of the fetus in the backed-up ultrasound target image of the fetus is marked as a region of interest, where a gray value of the region of interest is 1; and marking the region except the whole fetus body in the backed-up fetus ultrasonic target image as a non-interested region, wherein the gray value of the non-interested region is 0.
The label image obtained after the labeling processing of the embodiment is a binarized image. Exemplarily, fig. 5(a), 5(b) and 5(c) are label images corresponding to fig. 4(a), 4(b) and 4(c) which are three fetal ultrasound training images, respectively.
Referring to fig. 6, which is a flowchart of a method according to another embodiment of the present application, before the fetal ultrasound training image and the corresponding tag image are input into the preset image segmentation initial model in step S202, the method for automatically measuring a fetal head-hip diameter further includes:
s601, improving a preset Unet network to obtain an image segmentation initial model, wherein the image segmentation initial model comprises the following steps:
s6011, replacing a part of the plurality of convolutional layers in the downsampling process of the Unet network with expansion convolutional layers, wherein the expansion rates of the plurality of expansion convolutional layers are gradually increased along with the increase of the hierarchy;
s6012, setting batch normalization for each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network;
s6013, each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network are set to be connected through a residual error network.
In this embodiment, a model architecture diagram of an embodiment of the image segmentation initial model provided by the present application is shown in fig. 7, and is obtained by taking a U-net network as a basic idea and improving the U-net network. The image segmentation initial model of the embodiment mainly comprises a down-sampling part and an up-sampling part, wherein the down-sampling part is arranged on the left side in fig. 7, the up-sampling part is arranged on the right side, and the feature maps with the same resolution in the down-sampling part and the up-sampling part are subjected to feature fusion by adopting skip connection, so that a decoder can be helped to better recover details of a target. Referring to fig. 8, which is a parameter setting diagram of each layer in the image segmentation initial model provided in this embodiment, it should be noted that, for the size of the convolution kernel and the convolution step, the size may be specifically determined according to actual requirements, which is not limited herein.
In this embodiment, each convolution layer in the downsampling process includes two sub-convolution layers, and four convolution layers in the downsampling process are replaced with expanded convolution layers, wherein the expanded convolution rates are respectively set to be 3, 6, 12 and 18 along with the increase of the levels, so that the extraction capability of network context information can be improved on the premise of not increasing the number of parameters; note that the convolutional layer not using the extended convolution is a standard convolutional layer, that is, the extended convolution rate of the standard convolutional layer is 1. In this embodiment, the maximum pooling layer is adjacently connected after each expanded convolutional layer. In addition, each convolution layer in the down-sampling process comprises a sub-convolution layer, and a deconvolution mode is adopted.
Further, each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network are set with batch normalization, so that the robustness of the network can be improved, and the regularization strategy can be improved.
Further, as the depth of the network increases, the training error becomes larger and larger, and there is a network degradation phenomenon, which is not caused by over-fitting, so that a residual structure is added to each of the standard convolutional layer and the expanded convolutional layer, as shown in fig. 9, with the input of x, a residual function f (x), and the output of h (x) ═ f (x) + x.
Referring to fig. 10, a flowchart of an embodiment of a step S601 of the method for automatically measuring a hip diameter of a fetal head provided in the present application, that is, in step S601, a preset Unet network is modified to obtain an image segmentation initial model, and the method further includes:
s1001, setting a Leaky-relu activation function for each convolution layer and/or each expansion convolution layer in a down-sampling process and an up-sampling process of the Unet network;
s1002, adding a sigmoid activation function and a sigmoid cross entropy loss function to the last layer of the Unet network.
In this embodiment, a leak-relu activation function is set for each convolution layer and/or each expansion convolution layer in the downsampling process and the upsampling process of the Unet network, so that the gradient disappearance can be effectively avoided, and the linear function has strong calculation performance, and the leak-relu function is expressed as follows:
where a is a coefficient set to 0.2.
Further, an activation function added by the last layer of network of the Unet network is set as a sigmoid activation function, and a sigmoid cross entropy loss function is added; wherein the sigmoid function is expressed as follows:
furthermore, the training of the image segmentation initial model of the embodiment belongs to the binary classification in semantic segmentation, the fitting degree of the image segmentation initial model can be better evaluated by adopting a sigmoid cross entropy loss function, and the smaller the result of the evaluation index is, the better the result is;
since each in the sample image needs to be predictedThe classification of each pixel point is solved, so that the cross entropy of each pixel point is solved, the number of the pixel points in the sample image is N, and the real classification of each pixel is yiThe predicted probability is
The sigmoid cross entropy loss function is expressed as follows:
it should be noted that the loss value of one sample image can be solved by the equation (8-3), and the average value of the batch samples is required.
Referring to fig. 11, a flowchart of a method of an embodiment of step S202 provided in the present application, where the step S202 includes the following steps:
s1101, inputting the fetal ultrasound training image into an image segmentation initial model for training, outputting a prediction image, and calculating a loss value between the prediction image and a label image corresponding to the fetal ultrasound training image;
and S1102, based on the loss value, adopting a gradient descent algorithm to perform iterative training, minimizing a loss function, updating network parameters of the image segmentation initial model, wherein the network parameters comprise the weight and the offset of each layer of network, and taking the image segmentation initial model after the convergence training as an image segmentation target model.
In this embodiment, batch training is adopted, the batch size is set to be 1-10, the number of iterations is 100-300, and the batch size and the number of iterations can be specifically set according to the actual situation, which is not limited herein. When the iteration is carried out for the set times, the loss function can be converged to achieve the minimum global or local loss.
Optionally, when updating the network parameters of the image segmentation initial model, setting a learning rate η, and updating the network parameters by using an Adam optimizer, where an update formula is represented as follows:
where m is the number of iterations, L (θ) is the minimization loss function,
is the gradient of the parameter theta, theta comprises a weight w and a bias b, w is initialized to be a normal distribution, and b is initialized to be 0; in the present embodiment, the learning rate may be set to η 10-5。
Further, according to the formula (8-4), the updated weight w and the offset b are obtained:
wherein w(l)，b(l)Respectively the weight and the offset of the l-th layer,
are iteratively updated parameters.
Further, let the output of the ith layer of the image segmentation initial model be represented as y(l)＝w(l)×(x(l-1))Τ+b(l)It is possible to obtain:
wherein x is(l-1)Indicating the output of the l-1 tier network.
Order to
Equation (8-6) may become:
for bias b(l)By the same transformation, can obtain
It should be noted that the architecture and parameters of the well-trained image segmentation target model are stored in the database, and the model is saved in the format of h5 for facilitating later model calling. When the image segmentation target model is required to be called for prediction, the framework and the parameters of the image segmentation target model are called out to prepare for further prediction.
Referring to fig. 12, a flowchart of a method of an embodiment of step S204 provided in the present application, where the step S204 includes the following steps:
s1201, performing binarization processing on the predicted fetal image, wherein the binarization processing comprises correspondingly processing the pixel value of the predicted fetal image, which is greater than a preset threshold value, into 1, and correspondingly processing the pixel value of the predicted fetal image, which is less than or equal to the preset threshold value, into 0;
and S1202, fitting the predicted fetal image boundary after binarization processing, determining two farthest end points in the points with the pixel value of 1, and taking the distance between the two farthest end points as the length of the hip diameter of the fetal head.
In this embodiment, the image segmentation target model is used to perform image segmentation on the ultrasound image of the fetus to be tested, which is the ultrasound target image of the fetus in the test set. Exemplarily, three fetal ultrasound target images in the test set employed in the present embodiment are respectively illustrated with reference to fig. 13(a), 13(b), and 13 (c).
It should be noted that, since the predicted fetal image output by the image segmentation target model is not a binarized image, it is necessary to perform binarization processing on the image, in the binarization processing of the present embodiment, the preset threshold is determined to be 0.5, and fig. 14(a), 14(b), and 14(c) correspond to the predicted fetal images after prediction and binarization in fig. 13(a), 13(b), and 13(c), respectively.
Further, two farthest points with a pixel value of 1 are found from the predicted fetal image after binarization, the two points are two points of the fetal head-hip diameter, the length of the fetal head-hip diameter is calculated according to the pixel size between the two points, and fig. 15(a), 15(b), and 15(c) are respectively corresponding to fig. 13(a), 13(b), and 13(c) as measurement schematic diagrams of the fetal head-hip diameter.
According to the embodiment, the ultrasonic training image of the fetus and the corresponding label image are obtained, wherein the interested region of the label image is the whole body of the fetus, so that the whole body of the fetus can be conveniently trained and segmented by the initial model for segmenting the subsequent images; carrying out iterative convergence training on the image segmentation initial model to obtain a well-trained image segmentation target model, thereby facilitating image segmentation on the ultrasonic training image of the fetus by calling the model medically; the ultrasonic image of the fetus to be detected is input into the image segmentation target model to obtain a predicted fetus image, and the head-hip diameter length of the fetus can be rapidly and accurately calculated according to the more accurate shape of the whole body of the fetus in the predicted fetus image.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.
The embodiment also provides an automatic measuring system for the diameter of the head and the hip of the fetus, which is in one-to-one correspondence with the automatic measuring method for the diameter of the head and the hip of the fetus in the embodiment. As shown in fig. 16, the automatic fetal head-hip diameter measuring system includes an image acquiring module 1601, a training module 1602, a predicting module 1603, and a calculating module 1604. The functional modules are explained in detail as follows:
the image acquisition module 1601 is used for acquiring a fetal ultrasonic training image and a corresponding label image, wherein an interested area of the label image is a whole body of a fetus;
a training module 1602, configured to input the fetal ultrasound training image and the corresponding tag image into a preset image segmentation initial model, and perform iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
the prediction module 1603 is used for acquiring an ultrasonic image of the fetus to be detected, and inputting the ultrasonic image of the fetus to be detected into an image segmentation target model to obtain a predicted fetus image;
the calculating module 1604 is configured to calculate a fetal head-hip diameter length according to the predicted fetal image.
For specific definition of the automatic fetal head-hip diameter measurement system, reference may be made to the above definition of the automatic fetal head-hip diameter measurement method, and details are not repeated here. The modules in the automatic fetal head-hip diameter measuring system can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
As shown in fig. 17, based on the automatic measuring method for the head and hip diameters of the fetus, the application also provides a computer device, which may be a mobile terminal, a desktop computer, a notebook, a palm computer, a server, and other computing devices. The computer device comprises a processor 10, a memory 20 and a display 30. FIG. 17 shows only some of the components of a computer device, but it is to be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.
The storage 20 may in some embodiments be an internal storage unit of the computer device, such as a hard disk or a memory of the computer device. The memory 20 may also be an external storage device of the computer device in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), etc. provided on the computer device. Further, the memory 20 may also include both an internal storage unit and an external storage device of the computer device. The memory 20 is used for storing application software installed in the computer device and various data, such as program codes installed in the computer device. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 20 stores a computer program 40, and the computer program 40 can be executed by the processor 10 to implement the automatic fetal head-hip diameter measurement method according to the embodiments of the present application.
The processor 10 may be a Central Processing Unit (CPU), microprocessor or other data Processing chip in some embodiments, and is used for executing program codes stored in the memory 20 or Processing data, such as performing a fetal head and hip diameter automatic measurement method.
The display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch panel, or the like in some embodiments. The display 30 is used for displaying information at the computer device and for displaying a visual user interface. The components 10-30 of the computer device communicate with each other via a system bus.
In an embodiment, the following steps are implemented when the processor 10 executes the computer program 40 in the memory 20:
acquiring a fetus ultrasonic training image and a corresponding label image, wherein the region of interest of the label image is the whole body of the fetus;
inputting the ultrasonic training image of the fetus and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected into an image segmentation target model to obtain a predicted fetus image;
and calculating the length of the hip diameter of the head of the fetus according to the predicted fetus image.
The present embodiments also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring a fetus ultrasonic training image and a corresponding label image, wherein the region of interest of the label image is the whole body of the fetus;
inputting the ultrasonic training image of the fetus and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected into an image segmentation target model to obtain a predicted fetus image;
and calculating the length of the hip diameter of the head of the fetus according to the predicted fetus image.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above.
Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The above description is only for the preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application.
Claims (10)
1. A method for automatically measuring the hip diameter of a fetal head, which comprises the following steps:
acquiring a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is the whole body of the fetus;
inputting the fetal ultrasonic training image and the corresponding label image into a preset image segmentation initial model, and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
acquiring an ultrasonic image of a fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to the image segmentation target model to obtain a predicted fetus image;
and calculating the length of the hip diameter of the fetus head according to the predicted fetus image.
2. The method for automatically measuring the hip diameter of the fetus according to claim 1, wherein the acquiring of the ultrasonic training image and the corresponding tag image of the fetus comprises:
acquiring a plurality of fetal ultrasonic initial images, and shearing boundary interference areas of all the fetal ultrasonic initial images to obtain corresponding fetal ultrasonic target images;
performing data enhancement processing on all the fetal ultrasonic target images to obtain fetal ultrasonic training images;
backing up all the fetal ultrasonic target images, and marking a fetal whole body in the backed-up fetal ultrasonic target images as an interested area;
and performing data enhancement processing on all marked ultrasonic target images of the fetus to obtain corresponding label images.
3. The method for automatically measuring the hip diameter of the fetus according to claim 2, wherein the whole body of the fetus in the ultrasound target image of the fetus after backup is marked as a region of interest, wherein the gray value of the region of interest is 1; and marking the region except the whole body of the fetus in the backed-up fetus ultrasonic target image as a region of no interest, wherein the gray value of the region of no interest is 0.
4. The method for automatically measuring the hip diameter of the fetus according to claim 1, wherein before inputting the ultrasonic training image and the corresponding tag image into a preset image segmentation initial model, the method further comprises:
improving a preset Unet network to obtain the image segmentation initial model, including:
replacing part of a plurality of convolution layers in the downsampling process of the Unet network with expansion convolution layers, wherein the expansion rates of the plurality of expansion convolution layers are gradually increased along with the increase of the hierarchy;
setting batch normalization for each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network;
and setting each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network as residual network connection.
5. The method for automatically measuring the hip diameter of the fetus according to claim 4, wherein the improving the preset Unet network to obtain the initial image segmentation model further comprises:
setting a Leaky-relu activation function for each convolution layer and/or each expansion convolution layer in the down-sampling process and the up-sampling process of the Unet network;
and adding a sigmoid activation function and a sigmoid cross entropy loss function to the last layer of the Unet network.
6. The method for automatically measuring the hip diameter of the fetus according to claim 1, wherein the ultrasonic training image and the corresponding label image of the fetus are input into a preset image segmentation initial model, and the image segmentation initial model is subjected to iterative convergence training to obtain an image segmentation target model, comprising:
inputting the fetal ultrasonic training image into an image segmentation initial model for training, outputting a predicted image, and calculating a loss value between the predicted image and a label image corresponding to the fetal ultrasonic training image;
and based on the loss value, adopting a gradient descent algorithm to carry out iterative training, minimizing a loss function, updating network parameters of the image segmentation initial model, wherein the network parameters comprise the weight and the offset of each layer of network, and taking the image segmentation initial model after the convergence training as an image segmentation target model.
7. The automatic fetal head-hip diameter measurement method according to claim 1, wherein calculating the fetal head-hip diameter length according to the predicted fetal image comprises:
performing binarization processing on the predicted fetal image, wherein the corresponding processing of pixel values of the predicted fetal image, which are greater than a preset threshold value, into 1 and the corresponding processing of pixel values of the predicted fetal image, which are less than or equal to the preset threshold value, into 0;
and fitting the predicted fetal image boundary after the binarization processing, determining two farthest end points in the points with the pixel value of 1, and taking the distance between the two farthest end points as the length of the hip diameter of the fetal head.
8. An automatic fetal head-hip diameter measurement system, comprising:
the image acquisition module is used for acquiring a fetus ultrasonic training image and a corresponding label image, wherein the interested area of the label image is a whole fetus body;
the training module is used for inputting the fetal ultrasonic training image and the corresponding label image into a preset image segmentation initial model and performing iterative convergence training on the image segmentation initial model to obtain an image segmentation target model;
the prediction module is used for acquiring an ultrasonic image of the fetus to be detected, and inputting the ultrasonic image of the fetus to be detected to the image segmentation target model to obtain a predicted fetus image;
and the calculation module is used for calculating the length of the head and the hip diameter of the fetus according to the predicted fetus image.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method for automatic measurement of the hip diameter of a fetus as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, implements the steps of the method for automatic measurement of the hip diameter of a fetus as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110806990.8A CN113487581A (en) | 2021-07-16 | 2021-07-16 | Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110806990.8A CN113487581A (en) | 2021-07-16 | 2021-07-16 | Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113487581A true CN113487581A (en) | 2021-10-08 |
Family
ID=77939917
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110806990.8A Pending CN113487581A (en) | 2021-07-16 | 2021-07-16 | Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113487581A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114795258A (en) * | 2022-04-18 | 2022-07-29 | 浙江大学 | Child hip joint dysplasia diagnosis system |
| CN117618026A (en) * | 2023-11-27 | 2024-03-01 | 广东工业大学 | Fetal head circumference measurement method and system applied to ultrasonic image |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903292A (en) * | 2019-01-24 | 2019-06-18 | 西安交通大学 | A kind of three-dimensional image segmentation method and system based on full convolutional neural networks |
| CN110448335A (en) * | 2019-07-11 | 2019-11-15 | 暨南大学 | A kind of fetus head circumference full-automatic measuring method and device based on ultrasound image |
| CN110570350A (en) * | 2019-09-11 | 2019-12-13 | 深圳开立生物医疗科技股份有限公司 | two-dimensional follicle detection method and device, ultrasonic equipment and readable storage medium |
| CN111242929A (en) * | 2020-01-13 | 2020-06-05 | 中国科学技术大学 | Fetal skull shape parameter measuring method, system, equipment and medium |
| CN111739051A (en) * | 2020-06-02 | 2020-10-02 | 大连理工大学 | Multi-sequence MRI image segmentation method based on residual error network |
| CN111820948A (en) * | 2019-04-20 | 2020-10-27 | 无锡祥生医疗科技股份有限公司 | Fetal growth parameter measuring method and system and ultrasonic equipment |
| CN113034535A (en) * | 2019-12-24 | 2021-06-25 | 无锡祥生医疗科技股份有限公司 | Fetal head segmentation method, fetal head segmentation device and storage medium |
-
2021
- 2021-07-16 CN CN202110806990.8A patent/CN113487581A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903292A (en) * | 2019-01-24 | 2019-06-18 | 西安交通大学 | A kind of three-dimensional image segmentation method and system based on full convolutional neural networks |
| CN111820948A (en) * | 2019-04-20 | 2020-10-27 | 无锡祥生医疗科技股份有限公司 | Fetal growth parameter measuring method and system and ultrasonic equipment |
| WO2020215485A1 (en) * | 2019-04-20 | 2020-10-29 | 无锡祥生医疗科技股份有限公司 | Fetal growth parameter measurement method, system, and ultrasound device |
| CN110448335A (en) * | 2019-07-11 | 2019-11-15 | 暨南大学 | A kind of fetus head circumference full-automatic measuring method and device based on ultrasound image |
| CN110570350A (en) * | 2019-09-11 | 2019-12-13 | 深圳开立生物医疗科技股份有限公司 | two-dimensional follicle detection method and device, ultrasonic equipment and readable storage medium |
| CN113034535A (en) * | 2019-12-24 | 2021-06-25 | 无锡祥生医疗科技股份有限公司 | Fetal head segmentation method, fetal head segmentation device and storage medium |
| CN111242929A (en) * | 2020-01-13 | 2020-06-05 | 中国科学技术大学 | Fetal skull shape parameter measuring method, system, equipment and medium |
| CN111739051A (en) * | 2020-06-02 | 2020-10-02 | 大连理工大学 | Multi-sequence MRI image segmentation method based on residual error network |
Non-Patent Citations (1)
| Title |
|---|
| 李志昂 等: ""基于Ｄ－ＵＮＥＴ的胎儿产前超声检测"", 《四川大学学报（自然科学版）》, vol. 57, no. 4, 30 July 2020 (2020-07-30), pages 733 - 740 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114795258A (en) * | 2022-04-18 | 2022-07-29 | 浙江大学 | Child hip joint dysplasia diagnosis system |
| CN117618026A (en) * | 2023-11-27 | 2024-03-01 | 广东工业大学 | Fetal head circumference measurement method and system applied to ultrasonic image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110310256B (en) | Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium | |
| CN113487581A (en) | Method, system, equipment and storage medium for automatically measuring diameter of fetus head and buttocks | |
| CN109146891B (en) | Hippocampus segmentation method and device applied to MRI and electronic equipment | |
| US11684333B2 (en) | Medical image analyzing system and method thereof | |
| CN111415361B (en) | Method and device for estimating brain age of fetus and detecting abnormality based on deep learning | |
| CN111325266B (en) | Detection method and device for microcalcification clusters in breast molybdenum target image and electronic equipment | |
| CN111192678B (en) | Pathological microscopic image diagnosis and model training method, device, equipment and medium | |
| CN111583184A (en) | Image analysis method, network, computer device, and storage medium | |
| Qi et al. | Automatic lacunae localization in placental ultrasound images via layer aggregation | |
| CN114494215A (en) | Transformer-based thyroid nodule detection method | |
| CN110929730A (en) | Image processing method, image processing device, computer equipment and storage medium | |
| CN114066905A (en) | Medical image segmentation method, system and device based on deep learning | |
| CN111489318B (en) | Medical image enhancement method and computer-readable storage medium | |
| CN115908363B (en) | Tumor cell statistics method, device, equipment and storage medium | |
| CN111209946B (en) | Three-dimensional image processing method, image processing model training method and medium | |
| CN113822839B (en) | Medical image processing method, medical image processing device, computer equipment and storage medium | |
| CN115063637A (en) | Image classification method, storage medium, and program product | |
| CN114677503A (en) | Region-of-interest detection method and device, computer equipment and storage medium | |
| CN115760868A (en) | Colorectal and colorectal cancer segmentation method, system, device and medium based on topology perception | |
| CN115620913A (en) | Graph convolution network-based tumor recurrence prediction method after liver transplantation | |
| Yu et al. | CT segmentation of liver and tumors fused multi-scale features | |
| CN111539926B (en) | Image detection method and device | |
| CN115546089A (en) | Medical image segmentation method, pathological image processing method, device and equipment | |
| CN112426170A (en) | Placenta thickness determination method, device, equipment and storage medium | |
| Li et al. | Automatic detection of pituitary microadenoma from magnetic resonance imaging using deep learning algorithms |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211008 |