CN111369542B - Vessel marking method, image processing system, and storage medium - Google Patents
Vessel marking method, image processing system, and storage medium Download PDFInfo
- Publication number
- CN111369542B CN111369542B CN202010152105.4A CN202010152105A CN111369542B CN 111369542 B CN111369542 B CN 111369542B CN 202010152105 A CN202010152105 A CN 202010152105A CN 111369542 B CN111369542 B CN 111369542B
- Authority
- CN
- China
- Prior art keywords
- blood vessel
- point cloud
- network
- image
- cloud data
- 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
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10104—Positron emission tomography [PET]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The application relates to a blood vessel marking method, an image processing system and a storage medium. The method comprises the following steps: acquiring a medical image to be marked; inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image; performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point; and inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image. In the method, the obtained point cloud data are obtained by point clouding of the blood vessel segmentation result, the point cloud data can represent the whole space structure of the blood vessel region, and the point cloud data contain the space geometric information of each point, so that the space information of each point can be fully considered when the point cloud network carries out the marker classification, and the accuracy of the blood vessel marker result is greatly improved.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to a blood vessel labeling method, an image processing system, and a storage medium.
Background
CT Angiography (CTA) refers to a technology for developing and enhancing blood vessels after intravenous injection of an iodine-containing contrast agent, and the application of CTA to different parts of the body can obtain blood vessel enhanced images of different parts, and the blood vessel analysis process can be completed by identifying the CTA images. Vascular analysis plays an important role in the clinical field, mainly in terms of the process of segmenting a blood vessel portion from a CTA image and then assigning a label to the segmented blood vessel for distinguishing between different blood vessels.
Current methods of vessel analysis may include vessel reconstruction methods and deep learning methods: most of the vessel reconstruction methods are subtraction methods, gray values are subtracted based on CTA images and average images, the obtained image enhancement part is the vessel part, and then a doctor marks the vessel, but the method is time-consuming and labor-consuming, and the accuracy of marking results is low. With the continuous development of deep learning technology, the original CTA image is processed by using a deep learning network to complete the work of vessel segmentation, vessel marking and the like, but the deep learning network still cannot effectively mark the vessel.
In summary, the accuracy of vessel labeling using conventional techniques is low.
Disclosure of Invention
Based on this, it is necessary to provide a blood vessel marking method, an image processing system, and a storage medium, in order to solve the problem of low accuracy of blood vessel marking results in the conventional art.
A method of vessel marking, the method comprising:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
and inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
A vascular marking device, the device comprising:
the acquisition module is used for acquiring the medical image to be marked;
the image segmentation module is used for inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
the point clouding module is used for carrying out point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
the image marking module is used for inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
An image processing system comprises a medical imaging device and a computer device; the medical imaging device is used for acquiring medical images of the tested object and sending the medical images to the computer device;
the computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
and inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
And inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
The blood vessel marking method, the blood vessel marking device, the image processing system and the storage medium can firstly acquire the medical image to be marked, input the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image, and the blood vessel segmentation result can intuitively show a blood vessel region; then, carrying out point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result, wherein the point cloud data comprises the position coordinates of each point, so that the space geometric information of each point can be obtained; and finally, inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image. In the method, the obtained point cloud data are obtained by point clouding of the blood vessel segmentation result, the point cloud data can represent the whole space structure of the blood vessel region, and the point cloud data contain the space geometric information of each point, so that the space information of each point can be fully considered when the point cloud network carries out the marker classification, and the accuracy of the blood vessel marker result is greatly improved.
Drawings
FIG. 1 is a schematic diagram of an image processing system in one embodiment;
FIG. 2 is a flow chart of a method of vessel marking in one embodiment;
FIG. 2a is a schematic diagram of a vessel segmentation result in one embodiment;
FIG. 2b is a schematic diagram of a marker image in one embodiment;
FIG. 3 is a flow chart of a method of marking a blood vessel in another embodiment;
FIG. 3a is a general schematic of a vessel marking process in one embodiment;
FIG. 3b is a schematic diagram of a feature fusion module in one embodiment;
FIG. 3c is a flow chart of a method of marking a blood vessel in yet another embodiment;
FIG. 4 is a flow chart of a method of marking a blood vessel in yet another embodiment;
FIG. 4a is a schematic diagram of a first network and a second network process in one embodiment;
FIG. 5 is a flow diagram of a point cloud network training process in one embodiment;
FIG. 6 is a flow chart of a method of vessel marking in yet another embodiment;
FIG. 7 is a flow chart of a method of marking a blood vessel in yet another embodiment;
FIG. 8 is a block diagram of a vascular marking device in one embodiment;
fig. 9 is an internal structural diagram of a computer device in one embodiment.
Reference numerals illustrate:
11: medical imaging equipment; 12: a computer device.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The blood vessel marking method provided by the embodiment of the application can be applied to an image processing system shown in fig. 1. Wherein the system comprises a medical imaging device 11 and a computer device 12; the medical imaging device 11 may be an electronic computed tomography (Computed Tomography, CT) device, a magnetic resonance imaging (Magnetic Resonance Imaging, MRI) device, a positron emission computed tomography (Positron Emission Computed Tomography, PET) device or other medical device, for acquiring CT images, MRI images, PET images or medical images of other modalities of a patient, respectively, and transmitting the acquired medical images to the computer device 12. The computer device 12 is used to perform the steps in the method embodiments described below to achieve the process of accurate marking of blood vessels.
In one embodiment, a method for marking a blood vessel is provided, and the method is applied to the computer device in fig. 1 for example, and the embodiment relates to a specific process of segmenting, point clouding and marking a medical image to be marked by the computer device. As shown in fig. 2, the method comprises the steps of:
S101, acquiring a medical image to be marked.
Specifically, the medical image to be marked is obtained by scanning the object to be tested by the medical image equipment, and the medical image contains a blood vessel region, such as a CT angiography image. Alternatively, the medical image may be a coronary vessel image or a head and neck vessel image, wherein the head and neck and the coronary vessel contain a number of vessels, each having its own medical name, but not all vessels are of clinical interest to the clinician, such as the vessels of the head and neck vessel including aortic arch, brachiocephalic trunk, common carotid artery, internal carotid artery, vertebral artery, basilar artery, cerebral arterial ring, etc., which is often of interest, so our task will be to segment these vessels and label them.
Alternatively, the medical image may be obtained by the computer device directly from the medical imaging device; or the medical image equipment can be firstly transmitted to a post-processing workstation or an image archiving and communication system (Picture Archiving and Communication Systems, PACS), and then the medical image equipment can be acquired from the post-processing workstation or the PACS. Optionally, the computer device may acquire the medical images uploaded by the medical image device in real time, or may acquire all the medical images in the time period with a fixed time interval as a period. Optionally, the computer device may also acquire medical images to be marked from a hospital information management system (Hospital Information System, HIS), a clinical information system (Clinical Information System, CIS), a radiology information management system (Radiology Information System, RIS), an electronic medical record system (Electronic Medical Record, EMR), and a related medical image cloud storage platform.
Optionally, after the computer device acquires the medical image to be marked, the medical image may be preprocessed, including but not limited to image format conversion, window width and level setting, normalization, and the like.
S102, inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image.
Specifically, the computer device inputs the acquired medical image into a preset blood vessel segmentation network for segmentation processing, so as to obtain a blood vessel segmentation result of the medical image, wherein, as shown in fig. 2a, the blood vessel segmentation result can be a binarized mask image of a blood vessel region, that is, the separation of the blood vessel region and a background region is represented by different colors.
Alternatively, the vessel segmentation network may be a neural network, such as a convolutional neural network, a recurrent neural network, or the like, including but not limited to a segmentation network such as UNet, VNet, or the like. Optionally, the training manner of the vessel segmentation network may include: acquiring a certain number of training sample images, and marking the training sample images by a doctor with abundant experience to obtain a segmentation gold standard; inputting the training sample image into an initial vessel segmentation network to obtain a predicted vessel segmentation result; and calculating the loss between the predicted blood vessel segmentation result and the segmentation gold standard, and adjusting the network parameters of the initial blood vessel segmentation network by using the loss, thereby performing iterative training until the network converges.
S103, performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes position coordinates of each point.
Specifically, the point cloud is a point set expressing a target space under the same spatial reference system, and in this embodiment, the multi-label segmentation task of the blood vessel segmentation result is converted into a classification task based on the point set, so after the blood vessel segmentation result is obtained by the computer device, the blood vessel segmentation result can be subjected to point clouding, that is, the binarization mask image is subjected to point clouding. Optionally, the method may perform discrete sampling according to a preset sampling interval (spacing), and obtain position coordinates of sampling points, so as to obtain corresponding point cloud data, thereby obtaining space geometric information of each point. The position coordinates of each point in the point cloud data can be three-dimensional coordinates, and the computer equipment can extract the coordinate information of each point from the blood vessel segmentation result and record the coordinate information in a matrix form; each row in the matrix is a point on a blood vessel, each point has three columns corresponding to x, y and z axis coordinates, and the computer equipment can select a plurality of points from the matrix as point cloud data.
Optionally, the computer device may also resample the vessel segmentation results to a size of 1 x 1, and then, performing point clouding operation, so that the size range of the blood vessel segmentation result can be enlarged, and the selection range of the point cloud data is wider.
S104, inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
Specifically, the computer device inputs the point cloud data into a preset point cloud network, and can output the label of the sub-blood vessel to which each point belongs as a blood vessel marking result of the medical image, namely, the point cloud network can classify each point into different types according to the space geometric information of the point cloud network. Optionally, the point cloud network may further output a probability that each point belongs to a different sub-blood vessel, and then use a sub-blood vessel label corresponding to the maximum probability value as the label of the point.
Optionally, the point cloud network may be a PointNet network, and the training manner of the network may include: obtaining a certain number of training sample images, marking the training sample images by a doctor with abundant experience to obtain a blood vessel marking standard, dividing the training sample images by adopting the blood vessel dividing network to obtain sample blood vessel dividing results, and performing point clouding on the sample blood vessel dividing results to obtain corresponding sample point cloud data; inputting sample point cloud data into an initial point cloud network to obtain a predicted vessel marking result; calculating the loss between the predicted vessel marking result and the marking gold standard, adjusting the network parameters of the initial point cloud network by using the loss, and sequentially performing iterative training until the network converges.
Optionally, after obtaining the vessel marking result of the medical image, the computer device may also render it, and render a marking image with the same size as the medical image, and a schematic view of the marking image may be shown in fig. 2 b.
According to the blood vessel marking method provided by the embodiment, firstly, a medical image to be marked is acquired by computer equipment, the medical image is input into a preset blood vessel segmentation network, a blood vessel segmentation result of the medical image is obtained, and the blood vessel segmentation result can intuitively show a blood vessel region; then, carrying out point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result, wherein the computer equipment can acquire the space geometric information of each point because the point cloud data comprises the position coordinates of each point; and finally, inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image. In the method, the obtained point cloud data are obtained by point clouding of the blood vessel segmentation result, the point cloud data can represent the whole space structure of the blood vessel region, and the point cloud data contain the space geometric information of each point, so that the space information of each point can be fully considered when the point cloud network carries out the marker classification, and the accuracy of the blood vessel marker result is greatly improved.
In one embodiment, since semantic information of an image is also important for multi-label tasks, the semantic information can represent what object is in a certain area in the image, and when the neural network performs semantic analysis on the image, a feature map can contain high-level or low-level semantic information, and tasks such as image classification, semantic segmentation, object positioning and detection, target recognition and the like can be performed based on the semantic information. In order to further improve the accuracy of the obtained blood vessel marking result, the computer equipment can also comprehensively consider the semantic information of the point cloud data to determine the blood vessel marking result. Optionally, as shown in fig. 3, the method further includes:
s201, semantic information corresponding to the point cloud data is obtained from a feature layer of the blood vessel segmentation network.
Specifically, the semantic information corresponding to the point cloud data is derived from the blood vessel segmentation network, when the blood vessel segmentation network segments the medical image, the feature layer can extract the semantic information of each point of the medical image, the semantic information can represent which part in the image is a blood vessel, and then the semantic information of each point of the blood vessel region in the blood vessel segmentation result can also be obtained; and the point cloud data are obtained by carrying out point clouding on the blood vessel segmentation result, so that the computer equipment can also determine semantic information corresponding to the point cloud data. Optionally, the computer device may also obtain the above semantic information using other networks, including but not limited to self-encoders and the like.
S202, inputting the point cloud data and semantic information corresponding to the point cloud data into a point cloud network to obtain a blood vessel marking result.
Specifically, the computer device inputs the obtained point cloud data (including the position coordinates of each point) and semantic information corresponding to the point cloud data into the point cloud network, that is, the point cloud network synthesizes the space geometric information and the semantic information of each point, and further obtains a blood vessel marking result, so that the accuracy of the obtained blood vessel marking result can be further improved. An overall schematic of the vessel marking process in this embodiment can be seen in fig. 3 a.
Optionally, because the point cloud network integrates the space geometric information and semantic information of each point, the two kinds of information need to be feature fused, and the point cloud network further comprises a feature fusion module besides a series of convolution layers, full connection layers, pooling layers and the like. The feature fusion module comprises a feature layer structure of a blood vessel segmentation network and a terminal network structure of a traditional point cloud network, and is used for extracting features of semantic information and point cloud data respectively so as to perform feature fusion. As for the structural schematic diagram of the feature fusion module, see fig. 3b, optionally, as shown in fig. 3c, the step S202 may include:
S202a, inputting features of semantic information through a first feature module of a point cloud network and inputting features of point cloud data through a second feature module of the point cloud network;
s202b, fusing the features of the semantic information and the features of the point cloud data to obtain fused features; and carrying out convolution operation on the fusion characteristics through a convolution layer in the point cloud network to obtain a blood vessel marking result.
Specifically, a first feature module in the feature fusion module inputs features for extracting semantic information (the features of the semantic information may be provided by the above-mentioned vessel segmentation network), and a second feature module inputs features of point cloud data (the features of the point cloud data may be provided by a feature layer of the point cloud network). And then fusing the characteristics of the semantic information and the characteristics of the point cloud data to obtain fused characteristics, and carrying out convolution operation on the fused characteristics through convolution layers (including but not limited to two layers of convolution layers) in the point cloud network to obtain a blood vessel marking result. In this embodiment, the feature fusion module is located at the end position of the point cloud network, and optionally, other feature fusion methods may be used for fusion, which is not limited in this embodiment.
According to the blood vessel marking method provided by the embodiment, the computer equipment can also acquire semantic information corresponding to the point cloud data from the feature layer of the blood vessel segmentation network, and input the point cloud data and the semantic information corresponding to the point cloud data into the point cloud network, so that the point cloud network comprehensively considers the space geometric information and the semantic information of each point when performing classification marking, and the accuracy of a blood vessel marking result can be further improved.
In one embodiment, optionally, the above-mentioned vessel segmentation network may further include a first network and a second network, where the first network may be a coarse segmentation network and the second network may be a fine segmentation network, and the vessel segmentation result is obtained through two-step segmentation. As shown in fig. 4, referring to a specific process of inputting a medical image into a preset vessel segmentation network to obtain a vessel segmentation result of the medical image, S102 may include:
s301, inputting a medical image into a first network for global positioning to obtain a first segmentation result of the medical image; the first segmentation result comprises global information of the blood vessel.
Specifically, the computer device may input the medical image into the first network for global positioning, because there may be some blood vessels in the medical image that the doctor does not pay attention to, and then the approximate position of the blood vessel that the doctor pays attention to, that is, the global information of the blood vessel, may be positioned through the first network.
S302, inputting the first segmentation result and the medical image into a second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
Specifically, according to the first segmentation result (i.e. global information of the blood vessel), the computer device may acquire positioning information of the blood vessel region on the medical image, and then input the first segmentation result and the medical image into the second network at the same time to obtain local information of the blood vessel in the medical image, and optimize the first segmentation result to obtain a final required blood vessel segmentation result. Alternatively, the first network may employ a 4-time downsampled network structure, and the second network may employ a 2-time downsampled network structure to improve segmentation efficiency. It should be noted that, in this embodiment, global information and local partial information of a blood vessel may be obtained, and then the computer device may also obtain global semantic information and local semantic information of the blood vessel from the first network and the second network, respectively, and input the global semantic information and the local semantic information together with the point cloud data into the point cloud network. A schematic diagram of the first network and the second network processes can be seen in fig. 4 a.
Optionally, for better acquisition of global positioning information during training in the first network, the computer device may resample the training sample image to a size of 1 x 1 as input data, the blood vessel labeling can be expanded, so that the disappearance of the small blood vessel labeling caused by resampling is prevented, and the global information of the blood vessel is easier to obtain. The second network can ensure the 0.4×0.4×0.7 size of the medical image during training to ensure the segmentation accuracy. The first network and the second network may include an input layer, a standardization layer, an activation layer, a convolution layer, a pooling layer, a deconvolution layer, an output layer, and an interlayer connection, the loss function adopts an image region similarity measurement multi-label Dice loss function during network training, the network optimization function adopts an Adam adaptive optimizer, which can automatically adjust parameter update amplitude along with training state, and the training process of the network can refer to the method described in the above embodiments and is not repeated herein.
According to the blood vessel marking method provided by the embodiment, the computer equipment inputs the medical image into the first network for global positioning to obtain global information of blood vessels in the medical image, and then uses the second network to locally divide the medical image according to the global information to obtain a blood vessel division result of the medical image. The segmentation accuracy can be greatly improved by the segmentation mode of global information-local information, and the accuracy of the subsequent vessel marking result is further improved.
Optionally, in an embodiment, as shown in fig. 5, the training manner of the point cloud network for merging the point cloud data and the semantic information of the point cloud data may include:
s401, acquiring a sample image and a mark gold standard corresponding to the sample image;
s402, inputting a sample image into the blood vessel segmentation network to obtain a blood vessel segmentation result of the sample image;
s403, performing point clouding on a blood vessel segmentation result of the sample image to obtain sample point cloud data, and acquiring semantic information corresponding to the sample point cloud data from a feature layer of a blood vessel segmentation network;
s404, inputting sample point cloud data and semantic information corresponding to the sample point cloud data into an initial point cloud network to obtain a predicted blood vessel marking result;
s405, calculating the loss between the predicted blood vessel marking result and the marking gold standard, and training the initial point cloud network according to the loss to obtain the point cloud network.
The process of obtaining the sample image and the gold standard corresponding to the sample image, the process of dividing the sample image, the process of performing point clouding on the division result, and the process of outputting the predicted blood vessel marking result by the initial point cloud network may be described in the above embodiments, and the implementation principle is similar and will not be described herein. After the predicted blood vessel marking result is obtained in this embodiment, the computer device may calculate the loss between the predicted blood vessel marking result and the marking gold standard, and train the initial point cloud network according to the loss, to obtain a converged point cloud network. The training convergence point cloud network is adopted to carry out the vessel marking, so that the accuracy of the vessel marking result can be greatly improved.
In one embodiment, considering the size and limitation of the current GPU video memory, the whole medical image cannot be input into the blood vessel segmentation network for segmentation, the computer device may also perform the segmentation processing on the medical image first, and then input into the blood vessel segmentation network. Alternatively, as shown in fig. 6, S102 may include:
s501, sequentially acquiring a plurality of image blocks with preset sizes on a medical image.
Specifically, considering the size of the GPU video memory of the computer device, in order to avoid memory overflow, the medical image may be subjected to block processing. Wherein, the computer equipment can sequentially acquire image blocks with preset sizes from the medical images, and the preset sizes can be 128 3 。
S502, inputting each image block into a blood vessel segmentation network to obtain a blood vessel segmentation result of each image block.
Specifically, the computer device inputs each image block into the blood vessel segmentation model, and the blood vessel segmentation result of each image block can be obtained through the characteristic information of each image block.
S503, according to the acquisition sequence of each image block, the blood vessel segmentation result of each image block is spliced into the blood vessel segmentation result of the medical image.
Specifically, according to the acquisition order corresponding to each image block, the computer device may correspond the blood vessel segmentation result of each image block to the corresponding position of the original medical image, thereby stitching the blood vessel segmentation result of the whole medical image. In the stage of testing the blood vessel segmentation network, in order to avoid the problem of overflow of the GPU video memory, the computer equipment can also segment the training sample image by adopting a segmentation method so as to train.
According to the blood vessel marking method provided by the embodiment, the computer equipment inputs a plurality of image blocks into the blood vessel segmentation model, and the segmentation result is spliced into the blood vessel segmentation result of the whole medical image. Therefore, the overflow of the GPU video memory of the computer equipment can be reduced, the segmentation results are spliced according to the acquisition sequence corresponding to each image block, and the accuracy of the obtained blood vessel segmentation results can be improved.
For a better understanding of the overall process of the vessel marking method, which is described below by way of a complete embodiment, as shown in fig. 7, the method includes:
s601, acquiring a medical image to be marked;
s602, inputting a medical image into a first network to obtain global information, and obtaining a first segmentation result of the medical image;
s603, inputting the first segmentation result and the medical image into a second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
S604, performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result;
s605, semantic information corresponding to the point cloud data is obtained from the feature layer of the first network and the feature layer of the second network;
S606, inputting the point cloud data and semantic information corresponding to the point cloud data into a point cloud network to obtain a blood vessel marking result;
s607, performing image rendering on the blood vessel marking result to obtain a marking image with the same size as the medical image.
For the implementation process of each step, reference may be made to the description of the above embodiment, which is not repeated here.
It should be understood that, although the steps in the flowcharts of fig. 2 to 7 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps of fig. 2-7 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor does the order in which the sub-steps or stages are performed necessarily occur in sequence, but may be performed alternately or alternately with at least a portion of other steps or sub-steps or stages of other steps.
In one embodiment, as shown in fig. 8, there is provided a vascular marking device comprising: an acquisition module 21, an image segmentation module 22, a point clouding module 13 and an image marking module 14.
Specifically, the acquiring module 21 is configured to acquire a medical image to be marked;
the image segmentation module 22 is configured to input the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
the point clouding module 23 is configured to perform point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
the image marking module 24 is configured to input the point cloud data into a preset point cloud network, and obtain a blood vessel marking result of the medical image.
The blood vessel marking device provided in this embodiment may perform the above method embodiment, and its implementation principle and technical effects are similar, and will not be described herein.
In one embodiment, the obtaining module 21 is further configured to obtain semantic information corresponding to the point cloud data from a feature layer of the vessel segmentation network; the image marking module 24 is specifically configured to input the point cloud data and semantic information corresponding to the point cloud data into the point cloud network, and obtain a blood vessel marking result.
In one embodiment, the image marking module 24 is specifically configured to input features of semantic information through a first feature module of the point cloud network and input features of point cloud data through a second feature module of the point cloud network;
Fusing the characteristics of the semantic information and the characteristics of the point cloud data to obtain fused characteristics; and carrying out convolution operation on the fusion characteristics through a convolution layer in the point cloud network to obtain a blood vessel marking result.
In one embodiment, the vessel segmentation network includes a first network and a second network; the image segmentation module 22 is specifically configured to input the medical image into the first network for global positioning, so as to obtain a first segmentation result of the medical image; the first segmentation result includes global information of the blood vessel; and inputting the first segmentation result and the medical image into a second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
In one embodiment, the medical image includes a coronary vessel image or a head and neck vessel image.
In one embodiment, the apparatus further includes a training module, configured to obtain a sample image and a marker standard corresponding to the sample image; inputting the sample image into a blood vessel segmentation network to obtain a blood vessel segmentation result of the sample image; performing point clouding on a blood vessel segmentation result of the sample image to obtain sample point cloud data, and acquiring semantic information corresponding to the sample point cloud data from a feature layer of a blood vessel segmentation network; inputting sample point cloud data and semantic information corresponding to the sample point cloud data into an initial point cloud network to obtain a predicted blood vessel marking result; and calculating the loss between the predicted blood vessel marking result and the marking gold standard, and training the initial point cloud network according to the loss to obtain the point cloud network.
In one embodiment, the image segmentation module 22 is specifically configured to sequentially acquire a plurality of image blocks with preset sizes on the medical image; inputting each image block into a blood vessel segmentation network to obtain a blood vessel segmentation result of each image block; and splicing the blood vessel segmentation results of each image block into blood vessel segmentation results of the medical image according to the acquisition sequence of each image block.
In one embodiment, the apparatus further includes an image rendering module, configured to perform image rendering on the vessel marking result to obtain a marking image with the same size as the medical image.
The blood vessel marking device provided in this embodiment may perform the above method embodiment, and its implementation principle and technical effects are similar, and will not be described herein.
For specific limitations of the vascular marking device, reference is made to the above limitations of the vascular marking method, and no further description is given here. The various modules in the vascular marking device described above may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, as shown in FIG. 1, an image processing system is provided, comprising a medical imaging device 11 and a computer device 12;
the medical imaging device 11 is used for acquiring medical images of the tested object and sending the medical images to the computer device 12;
the computer device 12 comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
and inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring semantic information corresponding to the point cloud data from a feature layer of the blood vessel segmentation network;
and inputting the point cloud data and semantic information corresponding to the point cloud data into a point cloud network to obtain a blood vessel marking result.
In one embodiment, the processor when executing the computer program further performs the steps of:
Inputting features of semantic information through a first feature module of the point cloud network and inputting features of point cloud data through a second feature module of the point cloud network;
fusing the characteristics of the semantic information and the characteristics of the point cloud data to obtain fused characteristics; and carrying out convolution operation on the fusion characteristics through a convolution layer in the point cloud network to obtain a blood vessel marking result.
In one embodiment, the vessel segmentation network includes a first network and a second network; the processor when executing the computer program also implements the steps of:
inputting the medical image into a first network for global positioning to obtain a first segmentation result of the medical image; the first segmentation result includes global information of the blood vessel;
and inputting the first segmentation result and the medical image into a second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
In one embodiment, the medical image includes a coronary vessel image or a head and neck vessel image.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring a sample image and a mark gold standard corresponding to the sample image;
Inputting the sample image into a blood vessel segmentation network to obtain a blood vessel segmentation result of the sample image;
performing point clouding on a blood vessel segmentation result of the sample image to obtain sample point cloud data, and acquiring semantic information corresponding to the sample point cloud data from a feature layer of a blood vessel segmentation network;
inputting sample point cloud data and semantic information corresponding to the sample point cloud data into an initial point cloud network to obtain a predicted blood vessel marking result;
and calculating the loss between the predicted blood vessel marking result and the marking gold standard, and training the initial point cloud network according to the loss to obtain the point cloud network.
In one embodiment, the processor when executing the computer program further performs the steps of:
sequentially acquiring a plurality of image blocks with preset sizes on a medical image;
inputting each image block into a blood vessel segmentation network to obtain a blood vessel segmentation result of each image block;
and splicing the blood vessel segmentation results of each image block into blood vessel segmentation results of the medical image according to the acquisition sequence of each image block.
In one embodiment, the processor when executing the computer program further performs the steps of:
and performing image rendering on the blood vessel marking result to obtain a marking image with the same size as the medical image.
In one embodiment, the internal structure of the computer device may be as shown in fig. 9. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a vessel marking method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in fig. 9 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data includes a position coordinate of each point;
and inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
The computer readable storage medium provided in this embodiment has similar principles and technical effects to those of the above method embodiment, and will not be described herein.
In one embodiment, the computer program when executed by the processor further performs the steps of:
Acquiring semantic information corresponding to the point cloud data from a feature layer of the blood vessel segmentation network;
and inputting the point cloud data and semantic information corresponding to the point cloud data into a point cloud network to obtain a blood vessel marking result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
inputting features of semantic information through a first feature module of the point cloud network and inputting features of point cloud data through a second feature module of the point cloud network;
fusing the characteristics of the semantic information and the characteristics of the point cloud data to obtain fused characteristics; and carrying out convolution operation on the fusion characteristics through a convolution layer in the point cloud network to obtain a blood vessel marking result.
In one embodiment, the vessel segmentation network includes a first network and a second network; the computer program when executed by the processor also performs the steps of:
inputting the medical image into a first network for global positioning to obtain a first segmentation result of the medical image; the first segmentation result includes global information of the blood vessel;
and inputting the first segmentation result and the medical image into a second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
In one embodiment, the medical image includes a coronary vessel image or a head and neck vessel image.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a sample image and a mark gold standard corresponding to the sample image;
inputting the sample image into a blood vessel segmentation network to obtain a blood vessel segmentation result of the sample image;
performing point clouding on a blood vessel segmentation result of the sample image to obtain sample point cloud data, and acquiring semantic information corresponding to the sample point cloud data from a feature layer of a blood vessel segmentation network;
inputting sample point cloud data and semantic information corresponding to the sample point cloud data into an initial point cloud network to obtain a predicted blood vessel marking result;
and calculating the loss between the predicted blood vessel marking result and the marking gold standard, and training the initial point cloud network according to the loss to obtain the point cloud network.
In one embodiment, the computer program when executed by the processor further performs the steps of:
sequentially acquiring a plurality of image blocks with preset sizes on a medical image;
inputting each image block into a blood vessel segmentation network to obtain a blood vessel segmentation result of each image block;
and splicing the blood vessel segmentation results of each image block into blood vessel segmentation results of the medical image according to the acquisition sequence of each image block.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and performing image rendering on the blood vessel marking result to obtain a marking image with the same size as the medical image.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, or the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples merely represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (10)
1. A method of marking a blood vessel, the method comprising:
acquiring a medical image to be marked;
inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image;
performing point clouding on the blood vessel segmentation result to obtain point cloud data corresponding to the blood vessel segmentation result; the point cloud data comprises position coordinates of each point;
And inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image.
2. The method according to claim 1, wherein the method further comprises:
acquiring semantic information corresponding to the point cloud data from a feature layer of the vessel segmentation network;
inputting the point cloud data into a preset point cloud network to obtain a blood vessel marking result of the medical image, wherein the method comprises the following steps:
and inputting the point cloud data and semantic information corresponding to the point cloud data into the point cloud network to obtain the blood vessel marking result.
3. The method according to claim 2, wherein inputting the point cloud data and semantic information corresponding to the point cloud data into the point cloud network to obtain the blood vessel marking result includes:
inputting the characteristics of the semantic information through a first characteristic module of the point cloud network and inputting the characteristics of the point cloud data through a second characteristic module of the point cloud network;
fusing the characteristics of the semantic information and the characteristics of the point cloud data to obtain fused characteristics; and carrying out convolution operation on the fusion characteristics through a convolution layer in the point cloud network to obtain the vessel marking result.
4. A method according to any of claims 1-3, wherein the vessel segmentation network comprises a first network and a second network; inputting the medical image into a preset blood vessel segmentation network to obtain a blood vessel segmentation result of the medical image, wherein the blood vessel segmentation result comprises:
inputting the medical image into the first network for global positioning to obtain a first segmentation result of the medical image; the first segmentation result comprises global information of blood vessels;
inputting the first segmentation result and the medical image into the second network to acquire local information of the blood vessel, and optimizing the first segmentation result according to the local information of the blood vessel to acquire a blood vessel segmentation result of the medical image.
5. A method according to any one of claims 1-3, wherein the medical image comprises a coronary vessel image or a head and neck vessel image.
6. The method according to claim 2, wherein the training method of the point cloud network comprises:
acquiring a sample image and a mark gold standard corresponding to the sample image;
inputting the sample image into the blood vessel segmentation network to obtain a blood vessel segmentation result of the sample image;
Performing point clouding on a blood vessel segmentation result of the sample image to obtain sample point cloud data, and acquiring semantic information corresponding to the sample point cloud data from a feature layer of the blood vessel segmentation network;
inputting the sample point cloud data and semantic information corresponding to the sample point cloud data into an initial point cloud network to obtain a predicted blood vessel marking result;
and calculating the loss between the predicted blood vessel marking result and the marking gold standard, and training the initial point cloud network according to the loss to obtain the point cloud network.
7. The method according to claim 1, wherein inputting the medical image into a preset vessel segmentation network to obtain a vessel segmentation result of the medical image comprises:
sequentially acquiring a plurality of image blocks with preset sizes on the medical image;
inputting each image block into the blood vessel segmentation network to obtain a blood vessel segmentation result of each image block;
and splicing the blood vessel segmentation result of each image block into the blood vessel segmentation result of the medical image according to the acquisition sequence of each image block.
8. The method according to claim 1, wherein the method further comprises:
And performing image rendering on the blood vessel marking result to obtain a marking image with the same size as the medical image.
9. An image processing system, comprising a medical imaging device and a computer device;
the medical imaging device is used for acquiring medical images of the tested object and sending the medical images to the computer device;
the computer device comprising a memory storing a computer program and a processor implementing the steps of the method of any of claims 1-8 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010152105.4A CN111369542B (en) | 2020-03-06 | 2020-03-06 | Vessel marking method, image processing system, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010152105.4A CN111369542B (en) | 2020-03-06 | 2020-03-06 | Vessel marking method, image processing system, and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111369542A CN111369542A (en) | 2020-07-03 |
| CN111369542B true CN111369542B (en) | 2023-07-28 |
Family
ID=71211796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010152105.4A Active CN111369542B (en) | 2020-03-06 | 2020-03-06 | Vessel marking method, image processing system, and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111369542B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112348883B (en) * | 2020-11-24 | 2021-06-29 | 中国科学院自动化研究所 | Interventional instrument endpoint real-time positioning system, method and device in vascular interventional operation |
| CN112862833B (en) * | 2020-12-31 | 2024-07-26 | 上海联影智能医疗科技有限公司 | Blood vessel segmentation method, electronic device and storage medium |
| CN112862835A (en) * | 2021-01-19 | 2021-05-28 | 杭州深睿博联科技有限公司 | Coronary vessel segmentation method, device, equipment and computer readable storage medium |
| CN112801999B (en) * | 2021-02-05 | 2023-06-23 | 深圳睿心智能医疗科技有限公司 | Method and device for determining heart coronary artery dominance |
| CN113012146B (en) * | 2021-04-12 | 2023-10-24 | 东北大学 | Vascular information acquisition method and device, electronic equipment and storage medium |
| CN113763331A (en) * | 2021-08-17 | 2021-12-07 | 北京医准智能科技有限公司 | Coronary artery dominant type determination method, device, storage medium, and electronic apparatus |
| CN114037663A (en) * | 2021-10-27 | 2022-02-11 | 北京医准智能科技有限公司 | Blood vessel segmentation method, device and computer readable medium |
| CN114155193B (en) * | 2021-10-27 | 2022-07-26 | 北京医准智能科技有限公司 | Blood vessel segmentation method and device based on feature enhancement |
| CN115311407B (en) * | 2022-04-19 | 2023-09-12 | 北京和华瑞博医疗科技有限公司 | Feature point marking method, device, equipment and storage medium |
| CN114998273A (en) * | 2022-06-10 | 2022-09-02 | 深圳睿心智能医疗科技有限公司 | Blood vessel image processing method and device, electronic equipment and storage medium |
| CN118154616A (en) * | 2024-05-13 | 2024-06-07 | 瀚依科技(杭州)有限公司 | Blood vessel segmentation path extraction method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109858437A (en) * | 2019-01-30 | 2019-06-07 | 苏州大学 | Based on the baggage volume automatic classification method for generating inquiry network |
| CN110838122A (en) * | 2018-08-16 | 2020-02-25 | 北京大学 | Point cloud segmentation method and device and computer storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118334070A (en) * | 2017-05-09 | 2024-07-12 | 哈特弗罗公司 | System and method for anatomical segmentation in image analysis |
| CN110458112B (en) * | 2019-08-14 | 2020-11-20 | 上海眼控科技股份有限公司 | Vehicle detection method and device, computer equipment and readable storage medium |
| CN110490881A (en) * | 2019-08-19 | 2019-11-22 | 腾讯科技（深圳）有限公司 | Medical image dividing method, device, computer equipment and readable storage medium storing program for executing |
| CN110688947B (en) * | 2019-09-26 | 2022-06-28 | 西安知象光电科技有限公司 | Method for synchronously realizing human face three-dimensional point cloud feature point positioning and human face segmentation |
-
2020
- 2020-03-06 CN CN202010152105.4A patent/CN111369542B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110838122A (en) * | 2018-08-16 | 2020-02-25 | 北京大学 | Point cloud segmentation method and device and computer storage medium |
| CN109858437A (en) * | 2019-01-30 | 2019-06-07 | 苏州大学 | Based on the baggage volume automatic classification method for generating inquiry network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111369542A (en) | 2020-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111369542B (en) | Vessel marking method, image processing system, and storage medium | |
| CN111161270B (en) | Vascular segmentation method for medical image, computer device and readable storage medium | |
| CN111292314B (en) | Coronary artery segmentation method, device, image processing system and storage medium | |
| CN110914866B (en) | System and method for anatomical segmentation in image analysis | |
| CN111445449B (en) | Method, device, computer equipment and storage medium for classifying region of interest | |
| CN111325714B (en) | Method for processing region of interest, computer device and readable storage medium | |
| CN112465834B (en) | Blood vessel segmentation method and device | |
| CN111312374B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| US11715279B2 (en) | Weighted image generation apparatus, method, and program, determiner learning apparatus, method, and program, region extraction apparatus, method, and program, and determiner | |
| CN111951272A (en) | Method and device for segmenting brain image, computer equipment and readable storage medium | |
| CN113935943A (en) | Method, device, computer equipment and storage medium for intracranial aneurysm identification detection | |
| WO2021011775A1 (en) | Systems and methods for generating classifying and quantitative analysis reports of aneurysms from medical image data | |
| CN112634255B (en) | Method and device for establishing brain focus detection model and computer equipment | |
| CN111223158B (en) | Artifact correction method for heart coronary image and readable storage medium | |
| CN114732431B (en) | Computer-implemented method, apparatus, and medium for detecting vascular lesions | |
| CN113192031B (en) | Vascular analysis method, vascular analysis device, vascular analysis computer device, and vascular analysis storage medium | |
| CN113160199B (en) | Image recognition method and device, computer equipment and storage medium | |
| CN112419339B (en) | Medical image segmentation model training method and system | |
| CN111161240B (en) | Blood vessel classification method, apparatus, computer device, and readable storage medium | |
| CN114913174B (en) | Method, apparatus and storage medium for vascular system variation detection | |
| CN112530554B (en) | Scanning positioning method and device, storage medium and electronic equipment | |
| US20230368913A1 (en) | Uncertainty Estimation in Medical Imaging | |
| US12089976B2 (en) | Region correction apparatus, region correction method, and region correction program | |
| CN114972026A (en) | Image processing method and storage medium | |
| CN113902684A (en) | Image segmentation method and device, electronic device and storage medium |
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 |