CN110310256B - Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium - Google Patents
Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110310256B CN110310256B CN201910464256.0A CN201910464256A CN110310256B CN 110310256 B CN110310256 B CN 110310256B CN 201910464256 A CN201910464256 A CN 201910464256A CN 110310256 B CN110310256 B CN 110310256B
- Authority
- CN
- China
- Prior art keywords
- stenosis
- network
- image
- coronary artery
- coronary
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- 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
- 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
Abstract
According to the coronary stenosis detection method and device, the computer equipment and the storage medium, the obtained image to be detected is input into the coronary stenosis detection model, and a detection result is obtained; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network, wherein the output of the backbone network is respectively connected with the input of the segmentation network and the input of the stenosis analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result. The structure of the coronary artery stenosis detection model can show that the coronary artery stenosis detection model comprises a segmentation network and a stenosis analysis network, so that the coronary artery stenosis detection model is a multi-task model, can simultaneously realize segmentation of an input image to be detected and stenosis detection of the input image to be detected, and obtains detection results comprising at least two types, namely a coronary artery segmentation result and a coronary artery stenosis result.
Description
Technical Field
The present application relates to the field of medical detection technologies, and in particular, to a coronary stenosis detection method, apparatus, computer device, and storage medium.
Background
With the development of medical detection technology, how to quickly and accurately obtain a coronary artery detection result becomes a relatively concerned problem when performing index detection on various organs, for example, index detection on coronary arteries (coronary artery for short).
Currently, different types of coronary detection results can be obtained based on different models. Specifically, a Computed Tomography imaging (CTA) device may be used to obtain CTA image data of the coronary artery, further, a coronary artery segmentation model may be used to perform coronary artery segmentation on the obtained CTA image data to obtain a coronary artery segmentation result, and a coronary artery stenosis region positioning model may be used to perform stenosis region positioning on the obtained CTA image data to obtain a specific position of the coronary artery stenosis region.
However, the method for determining the coronary artery detection result has the problems of large memory occupation and low speed for obtaining the coronary artery detection result.
Disclosure of Invention
In a first aspect, a coronary stenosis detection method, the method comprising:
acquiring an image to be detected;
inputting an image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
In one embodiment, the stenosis analysis network comprises a classification network and a regression network; coronary stenosis results include the classification of the stenosis and the degree of the stenosis.
In one embodiment, the coronary segmentation result comprises a probability response map and a coronary segmentation image.
In one embodiment, the coronary stenosis detection model further comprises a pooling layer for outputting the region of interest image according to the probabilistic response map.
In one embodiment, the output of the pooling layer is connected to the inputs of the classification network and the regression network, respectively; the classification network is used for classifying the images of the region of interest to obtain narrow classifications; the regression network is used for carrying out narrow degree estimation on the images of the narrow region of interest to obtain the narrow degree.
In one embodiment, the backbone network is used to extract image features from the image to be detected.
In one embodiment, the coronary segmentation image includes the locations of keypoints; the method further comprises the following steps:
obtaining the position of the coronary artery central line in the image to be detected according to the probability response map;
and obtaining a coronary artery segmentation result according to the position of the central line of the coronary artery and the position of the key point.
In a second aspect, a coronary stenosis detection apparatus, the apparatus comprising:
the acquisition module is used for acquiring an image to be detected;
the detection module is used for inputting the image to be detected into the coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
In a third aspect, a computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the coronary stenosis detection method according to any of the embodiments of the first aspect when executing the computer program.
In a fourth aspect, a computer-readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the coronary stenosis detection method according to any one of the embodiments of the first aspect.
According to the coronary stenosis detection method and device, the computer equipment and the storage medium, the obtained image to be detected is input into the coronary stenosis detection model, and a detection result is obtained; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network, wherein the output of the backbone network is respectively connected with the input of the segmentation network and the input of the stenosis analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result. The structure of the coronary artery stenosis detection model can show that the coronary artery stenosis detection model comprises a segmentation network and a stenosis analysis network, so that the coronary artery stenosis detection model is a multi-task model, can simultaneously segment an input image to be detected and carry out stenosis detection on the input image to be detected to obtain detection results containing at least two types of results, namely a coronary artery segmentation result and a coronary artery stenosis result. In addition, the segmentation network and the narrow analysis network share parameters in the backbone network, so that the segmentation network and the narrow analysis network have strong correlation, and therefore, the learning performance of the segmentation network and the narrow analysis network can be improved in an auxiliary manner when the network structure is optimized, the risk of overfitting is reduced, the situation of practical application is better met, and the detection accuracy of a detection result is improved.
Drawings
FIG. 1 is a schematic diagram illustrating an internal structure of a computer device according to an embodiment;
FIG. 2 is a flow chart of a coronary stenosis detection method according to an embodiment;
FIG. 3 is a diagram of a network architecture, according to an embodiment;
FIG. 4 is a diagram of a network architecture provided in one embodiment;
FIG. 5 is a diagram of a network architecture provided in one embodiment;
FIG. 6 is a diagram of a network architecture provided by one embodiment;
FIG. 7 is a diagram of a network architecture, according to an embodiment;
FIG. 8 is a diagram of a network architecture provided by one embodiment;
FIG. 9 is a flow diagram of a training method provided by an embodiment;
FIG. 10 is a flow chart of a coronary stenosis detection method according to one embodiment;
fig. 11 is a schematic structural diagram of a coronary stenosis detection apparatus according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The coronary stenosis detecting method provided by the present application can be applied to a computer device as shown in fig. 1, where the computer device can be a terminal, and its internal structure diagram can be as shown in fig. 1. The computer device includes a processor, a memory, a network 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 comprises a nonvolatile 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 an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a coronary stenosis detection 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, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 1 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
The following describes in detail the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems by embodiments and with reference to the drawings. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
Fig. 2 is a flowchart of a coronary stenosis detection method according to an embodiment. The execution subject of the embodiment is the computer device shown in fig. 1, and the embodiment relates to a specific process of performing stenosis detection on an image to be detected by the computer device by using a coronary artery stenosis detection model. As shown in fig. 2, the method includes:
and S101, acquiring an image to be detected.
The image to be detected represents an image which needs to be detected currently, and is an image containing a coronary artery type structure, and specifically may include structures of a left coronary artery blood vessel, a right coronary artery blood vessel, or other tissues or organs adjacent to the coronary artery. The image to be detected may include, but is not limited to, a conventional CT image, a CTA image, an MRI image, a PET-MRI image, and the like, which is not limited in this embodiment. In practical application, the computer equipment can be connected with the scanning equipment to scan the coronary artery structure of the human body to obtain an image to be detected. Optionally, the computer device may also obtain the image to be detected containing the coronary artery structure directly from the database or from the internet by downloading, which is not limited in this embodiment.
S102, inputting an image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
The coronary artery stenosis detection model is used for performing coronary artery segmentation and coronary artery stenosis detection on an image to be detected so as to obtain a detection result, and the detection result can simultaneously comprise a coronary artery segmentation result and a coronary artery stenosis result. The coronary artery segmentation result can comprise various images describing the segmentation result, such as a segmentation image, a probability response map and the like; the coronary stenosis result may include various kinds of detection results or values describing coronary stenosis, such as indication information of stenosis classification, a position of a stenosis region, a size of the stenosis region, an estimation value of a stenosis degree, and the like.
The main network is used for extracting image features from the image to be detected, so that the feature image of the image to be detected is obtained. The backbone network may be a convolutional neural network, and may specifically include but not limited to a convolutional neural network such as V-NET, U-NET, VGG, ResNet, densneet, and the like, which is not limited in this embodiment. The backbone network in this embodiment is a V-NET network. The segmentation network is used for carrying out segmentation processing on the characteristic image output by the main network to obtain the coronary artery segmentation result. The stenosis analysis network is used for performing stenosis analysis on a coronary artery structure on the feature image output by the main network, so as to obtain the coronary artery stenosis result.
In addition, the coronary stenosis detection model in this embodiment may include a network structure as shown in fig. 3, which includes a backbone network, a segmentation network, and a stenosis analysis network, and an output of the backbone network is connected to inputs of the segmentation network and the stenosis analysis network, respectively.
In this embodiment, the coronary artery stenosis detection model shown in fig. 3 is used to realize segmentation and coronary artery stenosis detection of an image to be detected, and the specific process includes: when the computer device obtains the image to be detected based on the step of S101, the image to be detected may be input to the backbone network for feature extraction, so as to obtain a feature image of the detected image; the computer equipment further inputs the characteristic image into a segmentation network to carry out segmentation processing on the coronary artery structure to obtain a coronary artery segmentation result; meanwhile, the computer equipment can input the characteristic image into a stenosis analysis network to carry out coronary stenosis analysis, and a coronary stenosis result is obtained.
Optionally, the coronary artery stenosis detection model in this embodiment may further include a network structure as shown in fig. 4, where the network structure includes a backbone network, a segmentation network, and a stenosis analysis network, and an output of the backbone network is connected to inputs of the segmentation network and the stenosis analysis network, respectively; the output of the segmentation network is connected to the input of the stenosis analyzing network.
Specifically, the coronary artery stenosis detection model shown in fig. 4 is used to realize segmentation and coronary artery stenosis detection of an image to be detected, and the specific process includes: when the computer device obtains the image to be detected based on the step of S101, the image to be detected may be input to the backbone network for feature extraction, so as to obtain a feature image of the detected image; the computer equipment inputs the characteristic image into a segmentation network to carry out segmentation processing on the coronary artery structure, and a coronary artery segmentation result is obtained; further, the computer device can input the coronary artery segmentation result and the characteristic image into a stenosis analysis network to perform coronary artery stenosis analysis, so as to obtain a coronary artery stenosis result. Optionally, the computer device may further process the coronary artery segmentation result and the feature image to obtain an image on the feature image, where the image includes the coronary artery region to be detected, and then input the image of the coronary artery region to be detected to the stenosis analysis network to perform stenosis analysis on the coronary artery region, so as to obtain a coronary artery stenosis result.
In summary, in the coronary stenosis detection method provided in the above embodiment, the obtained image to be detected is input to the coronary stenosis detection model, so as to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network, wherein the output of the backbone network is respectively connected with the input of the segmentation network and the input of the stenosis analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result. The structure of the coronary artery stenosis detection model can show that the coronary artery stenosis detection model comprises a segmentation network and a stenosis analysis network, so that the coronary artery stenosis detection model is a multi-task model, can simultaneously segment an input image to be detected and carry out stenosis detection on the input image to be detected to obtain detection results containing at least two types of results, namely a coronary artery segmentation result and a coronary artery stenosis result. In addition, the segmentation network and the narrow analysis network share parameters in the backbone network, so that the segmentation network and the narrow analysis network have strong correlation, and therefore, the learning performance of the segmentation network and the narrow analysis network can be improved in an auxiliary manner when the network structure is optimized, the risk of overfitting is reduced, the situation of practical application is better met, and the detection accuracy of the detection result is improved.
The application also provides three network structures for realizing coronary artery stenosis detection of the image to be detected, and the method specifically comprises the following steps:
in a first application scenario, based on the network structure shown in fig. 3, the present application further provides a network structure, as shown in fig. 5, a narrow analysis network in the network structure includes a classification network and a regression network. In this application, the corresponding coronary stenosis results include a stenosis classification and a stenosis degree.
The stenosis classification is used to describe whether a coronary artery region included in the image to be detected belongs to a stenosis region, and may be specifically represented by characters, numbers, letters, and the like. The stenosis degree is used to describe the stenosis degree of the coronary artery region included in the image to be detected, and may be a specific numerical value or a percentage numerical value, and the magnitude of the numerical value represents the magnitude of the stenosis degree. The function of a classification network for classifying a specific coronary artery region in an input feature image and determining whether the coronary artery region belongs to a narrow region is explained on the basis of the network configuration shown in the embodiment of fig. 3, and if not, a classification result indicating "yes" is output, and if not, a classification result indicating "no" is output. Accordingly, the regression network is used to perform stenosis analysis on a specific coronary artery region in the input feature image, and obtain a stenosis degree of the coronary artery region, which is equivalent to an estimation value of the stenosis degree.
In the application scenario, the network structure shown in fig. 5 is used to segment or detect the image to be detected, and the specific process includes: inputting the obtained image to be detected into a backbone network by the computer equipment for feature extraction to obtain a feature image; then the characteristic image is further input into a segmentation network to carry out segmentation processing on the coronary artery structure, and a coronary artery segmentation result is obtained; meanwhile, the computer equipment can input the characteristic images into the classification network for classification processing to obtain the result of narrow classification, and input the characteristic images into the regression network for narrow degree estimation to obtain the estimated value of the narrow degree.
In a second application scenario, based on the network structure shown in fig. 4, the present application further provides a network structure, as shown in fig. 6, the narrow analysis network in the network structure includes a classification network and a regression network, and the output of the segmentation network is connected to the inputs of the classification network and the regression network. In this application, the corresponding coronary stenosis results include a stenosis classification and a stenosis degree.
In the application scenario, the network structure shown in fig. 6 is used to segment or detect the image to be detected, and the specific process includes: inputting the obtained image to be detected into a backbone network by the computer equipment for feature extraction to obtain a feature image; the computer equipment inputs the characteristic image into a segmentation network to carry out segmentation processing on the coronary artery structure, and a coronary artery segmentation result is obtained; then, on one hand, the computer equipment inputs the coronary artery segmentation result and the characteristic image into the classification network simultaneously for classification processing to obtain a stenosis classification result, and on the other hand, the computer equipment inputs the coronary artery segmentation result and the characteristic image into the regression network simultaneously for stenosis degree estimation to obtain an estimated value of the stenosis degree. Optionally, the computer device may further process the coronary artery segmentation result and the feature image to obtain an image of a region to be detected on the feature image, and then input the image of the region to be detected to the classification network and the regression network respectively to perform classification and stenosis degree detection on the coronary artery indicated by the region, so as to obtain a stenosis classification and stenosis degree correspondingly.
As can be seen from the above description, in any of the network structures shown in fig. 3-6, the segmentation network and the stenosis analysis network (or the classification network and the regression network) share the output characteristics of the backbone network, which indicates that the coronary artery segmentation and the coronary artery stenosis analysis are related to each other, and particularly, the output of the segmentation network in fig. 4 and 6 is connected to the input of the stenosis analysis network, so that the relationship of the stenosis analysis network is stronger, and therefore, when the network structure is optimized, the learning performance of each other can be improved in an auxiliary manner, the risk of overfitting easily generated in the multitask network structure is reduced, and the detection accuracy of the network structure is further improved. In practical applications, the coronary artery segmentation result optionally includes a probability response map and a coronary artery segmentation image. The coronary artery segmentation image represents an image after segmentation processing, the image contains a coronary artery structure, for example, the image can contain a left branch coronary artery blood vessel and a right branch coronary artery blood vessel after segmentation, and can also include an intersection point on the coronary artery blood vessel after segmentation, or other key points. The probability response graph is used for indicating the probability that each pixel point on the image to be detected or the characteristic image is expressed as coronary artery.
In a third application scenario, based on the network structure shown in fig. 6, the coronary stenosis detection model further includes a pooling layer, as shown in fig. 7. The pooling layer is used for outputting an interested area image according to the probability response graph, and the interested area image is a characteristic image containing all coronary structures or part of coronary structures. The function of the pooling layer for performing region-of-interest image extraction on the input feature image according to the region of interest to obtain a region-of-interest image is explained on the basis of the network structure shown in the embodiment of fig. 6.
Correspondingly, the output of the pooling layer is respectively connected with the input of the classification network and the input of the regression network; under the condition, the classification network is used for classifying the images of the interested region to obtain narrow classification; the regression network is used for estimating the stenosis degree of the region-of-interest images belonging to the stenosis, and an estimation value of the stenosis degree is obtained.
Under the above application, the network structure shown in fig. 7 is used to segment or examine the image to be examined, and the specific process includes: inputting the obtained image to be detected into a backbone network by the computer equipment for feature extraction to obtain a feature image; then, the characteristic image is further input into a segmentation network to carry out segmentation processing on the coronary artery structure, and a probability response graph is obtained; then the computer device obtains the interested region containing the coronary artery structure to be detected by analyzing the coronary artery probability value indicated by the probability response diagram, and then inputs the interested region and the characteristic image into the pooling layer to extract the image of the interested region, thus obtaining the image of the interested region containing the coronary artery structure to be detected. Then, the computer device can input the image of the region of interest into a regression network to perform stenosis degree estimation of the region, so as to obtain an estimation value of the stenosis degree, and input the image of the region of interest into a classification network to perform stenosis classification of the region, so as to identify whether the region belongs to stenosis, so as to obtain a stenosis classification result.
The network structure related to the embodiment uses the pooling layer to realize the feature extraction of the image blocks of the region of interest of the main network output feature image, so that the narrow analysis network connected with the network structure can pertinently perform narrow analysis on coronary structures contained in the image blocks of the region of interest, and the accurate determination of the narrow analysis is further improved.
In a fourth application scenario, based on the network structure shown in fig. 7, the coronary stenosis detection model further includes a convolutional layer, as shown in fig. 8. The function of the convolutional layer is added on the basis of the network structure shown in the embodiment of fig. 7, and the convolutional layer is used for performing depth feature extraction on the image of the region of interest to obtain a feature image of the region of interest. Correspondingly, the input of the convolution layer is connected with the pooling layer, and the output of the convolution layer is respectively connected with the input of the classification network and the input of the regression network; in this case, the classification network is used for classifying the characteristic images of the region of interest to obtain a narrow classification; the regression network is used for carrying out stenosis degree estimation on the characteristic image of the region-of-interest image belonging to stenosis to obtain an estimation value of the stenosis degree.
Under the above application, the network structure shown in fig. 8 is used to segment or examine the image to be examined, and the specific process includes: inputting the obtained image to be detected into a backbone network by the computer equipment for feature extraction to obtain a feature image; then, the characteristic image is further input into a segmentation network to carry out segmentation processing on the coronary artery structure, and a probability response graph is obtained; and then the computer equipment obtains an interested region containing the coronary artery structure to be detected by analyzing the coronary artery probability value indicated by the probability response diagram, and then inputs the interested region and the characteristic image into the pooling layer to extract an interested region image so as to obtain an interested region image containing the coronary artery structure to be detected. Then, inputting the region-of-interest image into a convolutional layer by computer equipment for depth feature extraction to obtain a feature image of the region-of-interest image; finally, the computer equipment inputs the characteristic image of the region of interest into a regression network to estimate the stenosis degree of the region, so as to obtain an estimation value of the stenosis degree, and inputs the characteristic image of the region of interest into a classification network to classify the stenosis of the region, so as to identify whether the region belongs to stenosis or not, so as to obtain a result of the stenosis classification.
The network structure related to the embodiment uses the convolutional layer to realize further depth feature extraction of the image blocks in the region of interest, so that the extracted features can better adapt to the analysis requirements of the narrow analysis network, and the relevance of the narrow analysis network, the segmentation network and the backbone network is enhanced, thereby improving the detection accuracy of the whole network structure.
As can be seen from the foregoing description, the coronary artery stenosis detection model is a network model trained by a computer device in advance, so the present application further provides a method for training the coronary artery stenosis detection model, fig. 9 is a flowchart of a training method provided in an embodiment, which relates to a process in which the computer device trains an initial coronary artery stenosis detection model according to a plurality of sample images and using gold standard images corresponding to the sample images as supervision information, as shown in fig. 9, where the process includes:
s201, obtaining a sample image.
The sample image represents an image currently used when training is needed, and is the same as the type of the image to be detected described in the foregoing S101, and specific contents may refer to the foregoing description, and redundant description is not repeated here.
S202, inputting the sample image into an initial coronary artery stenosis detection model for training by taking the gold standard image corresponding to the sample image as supervision information to obtain the coronary artery stenosis detection model.
The gold standard image is a marked image, and may be an image formed by marking various morphological structures included in the coronary artery structure on the sample image by a computer device in advance using different labels, for example, a left branch coronary artery, a right branch coronary artery, an intersection of the coronary arteries, and the like in the coronary artery structure. It should be noted that, when marking the intersection of the coronary artery, the intersection may be further expanded first, and the expanded intersection is marked with the same label, and if the coordinate value corresponding to the expanded intersection exceeds the range of the coronary artery blood vessel, the corresponding pixel point that exceeds is marked as the background. In addition, the computer device can mark the boundary of the coronary artery stenosis region on the sample image in advance, and estimate the stenosis degree of the marked stenosis region according to the cross-sectional area of the central line of the blood vessel, and optionally, the computer device can also estimate the stenosis degree of the marked stenosis region on the sample image by combining with the DSA diagnosis report. In addition to the above labeling method, the labeling method may further include labeling coronary vessels of different segments on the sample image using different labels, and labeling different intersection points or key points on the sample image using different labels, where the manner of labeling may be determined according to actual application requirements, and this embodiment is not limited to this embodiment. And after the computer equipment finishes all marking according to the method, taking the marked image as a gold standard image.
In this embodiment, when the computer device obtains a plurality of sample images and corresponding gold standard images, the plurality of sample images are input into the initial coronary artery stenosis detection model, the segmentation image and/or the stenosis detection image corresponding to the sample images are output, then parameters of the initial coronary artery stenosis detection model are adjusted according to a difference between the output segmentation image and/or the stenosis detection image and the gold standard image, and training is performed until a target loss function of the initial coronary artery stenosis detection model converges or the output segmentation image and/or the stenosis detection image is substantially consistent with the gold standard image corresponding to the input sample image, so as to obtain the trained coronary artery stenosis detection model, which is used in the detection process described in fig. 2.
In the training process, an objective loss function of the initial coronary stenosis detection model is involved, and the objective loss function adopts a multitask loss function, so that the objective loss function comprises a first loss function, a second loss function and a third loss function. Wherein, the first loss function can be obtained according to the output result of the segmentation network; the second loss function may be obtained from an output result of the classification network; the third loss function may be derived from the output of the regression network.
The target loss function in this embodiment is the sum of the weighted sums of the first, second, and third loss functions. Specifically, it can be expressed by the following relational expression (1):
Lz＝Ls+λ1*Lc+λ2*Lr (1)；
wherein Lz represents an objective loss function; ls represents a first loss function; lc represents a second loss function; lr represents a third loss function; lambda [ alpha ]1Representing a classification weight; lambda [ alpha ]2The regression weights are represented.
There is also an application scenario that, when the locations of the key points are included in the coronary artery segmentation image, after the computer device obtains the coronary artery segmentation result including the probability response map and the coronary artery segmentation image and the coronary artery stenosis result including the stenosis classification and the stenosis degree through the above embodiment, the computer device may further perform the following steps to obtain the location of the coronary artery centerline and the coronary artery segmentation result, as shown in fig. 10, the steps including:
s301, obtaining the position of the coronary artery central line in the image to be detected according to the probability response map.
In this embodiment, when the computer device obtains the probability response map, it may further select, from a probability response curve or a curved surface describing the probability response curve in each cross-sectional area of the blood vessel in the probability response map, a point with the highest probability response in each cross-sectional area as a control point, and sequentially connect the control points, where the connected curve is a partial coronary centerline. Then, the computer device selects one end on the coronary centerline as an initial point, and performs line segment tracking in the peripheral radius area through the adjacent relation or by taking the end point as a central point to supplement or extend the tail end of the coronary centerline, thereby obtaining the complete coronary centerline. And then the computer equipment can also carry out smoothing treatment on the obtained coronary artery central line by using a smoothing algorithm so as to eliminate abnormal control points.
S302, obtaining a coronary artery segmentation result according to the position of the central line of the coronary artery and the position of the key point.
When the computer device acquires the coronary artery central line, namely the position of the coronary artery central line is acquired, then the computer device can perform segmentation processing on the coronary artery central line according to the acquired positions of the key points, and specifically, a blood vessel indicated by a line segment between two key points is used as a segmented coronary artery blood vessel. After the multiple segments of blood vessels are obtained according to the method, the coronary artery segmentation result of the coronary artery structure is obtained.
It should be understood that although the steps in the flow charts of fig. 2, and fig. 9 and 10 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps of fig. 2, and fig. 9 and 10 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential.
In one embodiment, as shown in fig. 11, there is provided a coronary stenosis detection apparatus comprising: an acquisition module 11 and a detection module 12, wherein:
the acquisition module 11 is used for acquiring an image to be detected;
the detection module 12 is used for inputting the image to be detected into the coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
For specific definition of coronary stenosis detection, see the above definition of a coronary stenosis detection method, which is not described herein again. The modules in the coronary stenosis detecting apparatus may be implemented in whole or in part by software, hardware, or 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.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring an image to be detected;
inputting an image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, the computer program, when executed by a processor, further implementing the steps of:
acquiring an image to be detected;
inputting an image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the narrow analysis network; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result.
The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.
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, among others. 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 (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A coronary stenosis detection method, comprising: acquiring an image to be detected;
inputting the image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network, wherein the stenosis analysis network comprises a classification network and a regression network; the output of the backbone network is respectively connected with the input of the segmentation network and the input of the stenosis analysis network, the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result, the classification network is used for classifying the output result of the backbone network to obtain a stenosis classification, and the regression network is used for carrying out stenosis degree estimation on the output result of the backbone network to obtain a stenosis degree;
the coronary stenosis result comprises the stenosis classification and the stenosis degree; the stenosis classification is used for describing whether a coronary artery region contained in the image to be detected belongs to a stenosis region, and the stenosis degree is used for describing the stenosis degree of the coronary artery region contained in the image to be detected.
2. The method of claim 1, wherein the output of the segmentation network is connected to the inputs of the classification network and the regression network, the classification network is configured to classify the output of the segmentation network and the output of the backbone network to obtain the stenosis degree, and the regression network is configured to classify the output of the segmentation network and the output of the backbone network to obtain the stenosis degree.
3. The method of claim 1, wherein the coronary segmentation result comprises a probability response map and a coronary segmentation image.
4. The method of claim 3, wherein the coronary stenosis detection model further comprises a pooling layer for outputting a region of interest image from the probabilistic response map.
5. The method of claim 4, wherein the outputs of the pooling layer are connected to the inputs of the classification network and the regression network, respectively; the classification network is used for classifying the images of the region of interest to obtain the stenosis classification; the regression network is used for carrying out narrow degree estimation on the images of the narrow region of interest to obtain the narrow degree.
6. The method of claim 1, wherein the backbone network is used to extract image features from the image to be detected.
7. The method according to claim 3 or 4, wherein the coronary segmentation image comprises the locations of keypoints; the method further comprises the following steps:
obtaining the position of the coronary artery central line in the image to be detected according to the probability response map;
and obtaining a coronary artery segmentation result according to the position of the coronary artery central line and the position of the key point.
8. A coronary stenosis detection apparatus, comprising:
the acquisition module is used for acquiring an image to be detected;
the detection module is used for inputting the image to be detected into a coronary stenosis detection model to obtain a detection result; the coronary artery stenosis detection model comprises a backbone network, a segmentation network and a stenosis analysis network, wherein the stenosis analysis network comprises a classification network and a regression network; the output of the backbone network is connected with the input of the segmentation network and the input of the stenosis analysis network respectively; the detection result comprises a coronary artery segmentation result and a coronary artery stenosis result, the classification network is used for classifying the output result of the main network to obtain a stenosis classification, and the regression network is used for estimating the stenosis degree of the output result of the main network to obtain the stenosis degree; the coronary stenosis result comprises the stenosis classification and the stenosis degree; the stenosis classification is used for describing whether a coronary artery region contained in the image to be detected belongs to a stenosis region, and the stenosis degree is used for describing the stenosis degree of the coronary artery region contained in the image to be detected.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910464256.0A CN110310256B (en) | 2019-05-30 | 2019-05-30 | Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910464256.0A CN110310256B (en) | 2019-05-30 | 2019-05-30 | Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110310256A CN110310256A (en) | 2019-10-08 |
| CN110310256B true CN110310256B (en) | 2021-09-21 |
Family
ID=68075665
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910464256.0A Active CN110310256B (en) | 2019-05-30 | 2019-05-30 | Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110310256B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866931B (en) * | 2019-11-18 | 2022-11-01 | 东声(苏州)智能科技有限公司 | Image segmentation model training method and classification-based enhanced image segmentation method |
| CN110910370B (en) * | 2019-11-21 | 2023-02-28 | 北京理工大学 | CTA image coronary stenosis detection method and device |
| CN111012377B (en) * | 2019-12-06 | 2020-11-03 | 北京安德医智科技有限公司 | Echocardiogram heart parameter calculation and myocardial strain measurement method and device |
| CN111383259B (en) * | 2020-03-02 | 2023-08-22 | 联影智能医疗科技(北京)有限公司 | Image analysis method, computer device, and storage medium |
| CN111369528B (en) * | 2020-03-03 | 2022-09-09 | 重庆理工大学 | Coronary artery angiography image stenosis region marking method based on deep convolutional network |
| CN111292314A (en) * | 2020-03-03 | 2020-06-16 | 上海联影智能医疗科技有限公司 | Coronary artery segmentation method, coronary artery segmentation device, image processing system and storage medium |
| CN111340794B (en) * | 2020-03-09 | 2023-07-04 | 中山大学 | Quantification method and device for coronary artery stenosis |
| CN111445449B (en) * | 2020-03-19 | 2024-03-01 | 上海联影智能医疗科技有限公司 | Method, device, computer equipment and storage medium for classifying region of interest |
| CN111667456B (en) * | 2020-04-28 | 2023-11-14 | 北京理工大学 | Method and device for detecting vascular stenosis in coronary artery X-ray sequence radiography |
| CN112967284A (en) * | 2021-04-30 | 2021-06-15 | 平安科技（深圳）有限公司 | Blood vessel image segmentation analysis method, device, equipment and storage medium |
| CN114972221B (en) * | 2022-05-13 | 2022-12-23 | 北京医准智能科技有限公司 | Image processing method and device, electronic equipment and readable storage medium |
| CN114757944B (en) * | 2022-06-13 | 2022-08-16 | 深圳科亚医疗科技有限公司 | Blood vessel image analysis method and device and storage medium |
| CN116616804B (en) * | 2023-07-25 | 2023-10-13 | 杭州脉流科技有限公司 | Method, device, equipment and storage medium for acquiring intracranial arterial stenosis evaluation parameters |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446662A (en) * | 2018-04-02 | 2018-08-24 | 电子科技大学 | A kind of pedestrian detection method based on semantic segmentation information |
| CN109272510A (en) * | 2018-07-24 | 2019-01-25 | 清华大学 | The dividing method of tubular structure in a kind of 3 d medical images |
| CN109410267A (en) * | 2018-09-30 | 2019-03-01 | 数坤（北京）网络科技有限公司 | A kind of coronary artery segmentation appraisal procedure and system |
| CN109448005A (en) * | 2018-10-31 | 2019-03-08 | 数坤（北京）网络科技有限公司 | One kind being used for network model dividing method coronarius and equipment |
| CN109544566A (en) * | 2018-11-29 | 2019-03-29 | 上海联影医疗科技有限公司 | Coronary artery image partition method, device, computer equipment and storage medium |
| CN109686426A (en) * | 2018-12-29 | 2019-04-26 | 上海商汤智能科技有限公司 | Medical imaging processing method and processing device, electronic equipment and storage medium |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9700219B2 (en) * | 2013-10-17 | 2017-07-11 | Siemens Healthcare Gmbh | Method and system for machine learning based assessment of fractional flow reserve |
| CN108171698B (en) * | 2018-02-12 | 2020-06-09 | 数坤（北京）网络科技有限公司 | Method for automatically detecting human heart coronary calcified plaque |
| CN109222980A (en) * | 2018-06-19 | 2019-01-18 | 北京红云智胜科技有限公司 | Method of the measurement coronarogram based on deep learning as blood vessel diameter |
-
2019
- 2019-05-30 CN CN201910464256.0A patent/CN110310256B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446662A (en) * | 2018-04-02 | 2018-08-24 | 电子科技大学 | A kind of pedestrian detection method based on semantic segmentation information |
| CN109272510A (en) * | 2018-07-24 | 2019-01-25 | 清华大学 | The dividing method of tubular structure in a kind of 3 d medical images |
| CN109410267A (en) * | 2018-09-30 | 2019-03-01 | 数坤（北京）网络科技有限公司 | A kind of coronary artery segmentation appraisal procedure and system |
| CN109448005A (en) * | 2018-10-31 | 2019-03-08 | 数坤（北京）网络科技有限公司 | One kind being used for network model dividing method coronarius and equipment |
| CN109544566A (en) * | 2018-11-29 | 2019-03-29 | 上海联影医疗科技有限公司 | Coronary artery image partition method, device, computer equipment and storage medium |
| CN109686426A (en) * | 2018-12-29 | 2019-04-26 | 上海商汤智能科技有限公司 | Medical imaging processing method and processing device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110310256A (en) | 2019-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110310256B (en) | Coronary stenosis detection method, coronary stenosis detection device, computer equipment and storage medium | |
| CN109993726B (en) | Medical image detection method, device, equipment and storage medium | |
| CN110766730B (en) | Image registration and follow-up evaluation method, storage medium and computer equipment | |
| CN111080573B (en) | Rib image detection method, computer device and storage medium | |
| CN108846829B (en) | Lesion site recognition device, computer device, and readable storage medium | |
| CN110415792B (en) | Image detection method, image detection device, computer equipment and storage medium | |
| CN112308846B (en) | Blood vessel segmentation method and device and electronic equipment | |
| CN110570407B (en) | Image processing method, storage medium, and computer device | |
| CN110210519B (en) | Classification method, computer device, and storage medium | |
| CN110738702B (en) | Three-dimensional ultrasonic image processing method, device, equipment and storage medium | |
| CN110490841B (en) | Computer-aided image analysis method, computer device and storage medium | |
| CN110244249B (en) | Magnetic resonance scanning method, magnetic resonance scanning device, medical scanning equipment and storage medium | |
| CN115018863A (en) | Image segmentation method and device based on deep learning | |
| CN114387464A (en) | Vulnerable plaque identification method based on IVUS image, computer device, readable storage medium and program product | |
| CN111160441B (en) | Classification method, computer device, and storage medium | |
| CN110310314B (en) | Image registration method and device, computer equipment and storage medium | |
| CN114693671B (en) | Lung nodule semi-automatic segmentation method, device, equipment and medium based on deep learning | |
| CN116993812A (en) | Coronary vessel centerline extraction method, device, equipment and storage medium | |
| CN113160199B (en) | Image recognition method and device, computer equipment and storage medium | |
| CN113935943A (en) | Method, device, computer equipment and storage medium for intracranial aneurysm identification detection | |
| CN110766653B (en) | Image segmentation method and device, computer equipment and storage medium | |
| CN113362350A (en) | Segmentation method and device for cancer medical record image, terminal device and storage medium | |
| CN116681716B (en) | Method, device, equipment and storage medium for dividing intracranial vascular region of interest | |
| CN111145152A (en) | Image detection method, computer device, and storage medium | |
| CN114708259B (en) | CTA (computed tomography angiography) -based head and neck vascular stenosis detection method, device, equipment and 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 |