A data augmentation method, device, electronic device and storage medium

Abstract

The present invention discloses a data augmentation method, device, electronic device and storage medium. The method includes obtaining an image to be processed, and extracting the blood vessel centerline in the image to be processed; determining endpoints in the image to be processed based on the blood vessel centerline; on the path between any two of the endpoints, determining an anchor point, and performing local transformation on all points on the path based on the anchor point. By performing local transformation on the blood vessel path between any two endpoints in the image to be processed based on the anchor point, the present invention realizes data augmentation of the image to be processed, solves the problems of poor training data quality and insufficient training data, improves the quality of training data, and thus enhances the generalization ability of the model.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular, to a data enhancement method, apparatus, electronic device, and storage medium. Background Art

[0002] Computed Tomography (CT) angiography uses multi-slice spiral CT reconstruction technology. By intravenously injecting a contrast agent and rapidly completing a scan of a specific range when the contrast agent concentration in the blood vessels reaches its peak, the blood vessel images of various parts of the body can be displayed through deep learning algorithms. Deep learning algorithms are a type of statistical algorithm for analyzing data features. Therefore, a large amount of data is required to train the network framework. However, due to the particularity of medical data, it is difficult to obtain raw CT data, and professional data annotators are needed to label the tags. Therefore, it is difficult to obtain a large amount of CT data.

[0003] Currently, the way to obtain CT data is to perform blood vessel staining on CT images through a blood vessel staining scheme, and use the CT images after blood vessel staining as the training data for deep learning. However, although the above method solves the problem of labeling training data, there are still problems of poor quality and insufficient amount of training data. Summary of the Invention

[0004] The present invention provides a data enhancement method, apparatus, electronic device, and storage medium to solve the problems of poor quality and insufficient amount of training data.

[0005] According to one aspect of the present invention, a data enhancement method is provided, including:

[0006] Obtain an image to be processed, and extract the blood vessel centerline in the image to be processed;

[0007] Determine the endpoints in the image to be processed based on the blood vessel centerline;

[0008] On the path between any two of the endpoints, determine an anchor point, and perform a local transformation on all points on the path based on the anchor point.

[0009] According to another aspect of the present invention, a data enhancement apparatus is provided, including:

[0010] A centerline extraction module, configured to obtain an image to be processed and extract the blood vessel centerline in the image to be processed;

[0011] An endpoint determination module, configured to determine the endpoints in the image to be processed based on the blood vessel centerline;

[0012] A local transformation module, configured to determine an anchor point on a path between any two of the endpoints, and perform local transformation on all points on the path based on the anchor point.

[0013] According to another aspect of the present invention, there is provided an electronic device, including:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the at least one processor is enabled to execute the data enhancement method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, there is provided a computer-readable storage medium storing computer instructions for implementing the data enhancement method according to any embodiment of the present invention when executed by a processor.

[0018] The technical solution of the embodiment of the present invention obtains a to-be-processed image, extracts the blood vessel centerline in the to-be-processed image; determines endpoints in the to-be-processed image based on the blood vessel centerline; determines an anchor point on a path between any two endpoints, and performs local transformation on all points on the path based on the anchor point. By performing local transformation on the blood vessel path between any two endpoints in the to-be-processed image based on the anchor point, data enhancement of the to-be-processed image is realized, the problems of poor training data quality and insufficient training data are solved, the quality of training data is improved, and thus the generalization ability of the model is enhanced.

[0019] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0021] Figure 1 is a flowchart of a data enhancement method provided in Embodiment 1 of the present invention;

[0022] Figure 2It is a schematic diagram of local rotation based on anchor points provided in the first embodiment of the present invention;

[0023] Figure 3 It is a flowchart of a data enhancement method provided in the second embodiment of the present invention;

[0024] Figure 4 It is a schematic structural diagram of a data enhancement device provided in the third embodiment of the present invention;

[0025] Figure 5 It is a schematic structural diagram of an electronic device provided in the fourth embodiment of the present invention. Detailed implementation manners

[0026] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0028] Figure 1 It is a flowchart of a data enhancement method provided in the first embodiment of the present invention. This embodiment is applicable to the situation of data enhancement for the vascular mask image after segmentation based on the vascular segmentation scheme. This method can be executed by a data enhancement device, which can be implemented in the form of hardware and / or software, and the data enhancement device can be configured in the electronic device provided in the embodiment of the present invention. As Figure 1 shown, the method includes:

[0029] S110. Obtain the image to be processed, and extract the vascular centerline in the image to be processed.

[0030] Among them, the image to be processed refers to the segmented vascular mask image. Specifically, image segmentation can be performed on the CT image based on segmentation schemes such as ResUnet, graph neural network, and sparse convolution based on point cloud (SP-Conv) to obtain the vascular mask image, and the vascular mask image is used as the image to be processed; Exemplarily, the image to be processed includes but is not limited to coronary artery segmentation images, pulmonary vascular arteriovenous segmentation images, etc., which are not limited here. In this embodiment, a segmented mask image is obtained as the image to be processed, and the vascular centerline in the image to be processed is extracted.

[0031] Based on the above embodiment, optionally, the extracting the vascular centerline in the image to be processed includes: extracting the vascular skeleton of the image to be processed based on a skeleton extraction algorithm to obtain a skeleton image; determining the vascular centerline in the image to be processed based on the skeleton image.

[0032] In this embodiment, the extraction method of the vascular centerline is specifically as follows: First, the vascular skeleton in the image to be processed can be extracted according to the skeleton extraction algorithm to obtain a skeleton image; Among them, the skeleton extraction algorithm includes but is not limited to the Skeletonize algorithm, which is not limited here; Then, the path between the starting point and the ending point of the vascular skeleton can be extracted in the skeleton image as the vascular centerline.

[0033] S120. Determine the endpoints in the image to be processed based on the vascular centerline.

[0034] Among them, the endpoints refer to the endpoints of the blood vessels in the image to be processed. In this embodiment, all points on the vascular centerline are traversed, and it is judged whether they are endpoints. If they are endpoints, the endpoint coordinates are recorded.

[0035] Based on the above embodiment, optionally, the determining the endpoints in the image to be processed based on the vascular centerline includes: traversing the points on the vascular centerline, and performing endpoint detection on the points on the vascular centerline to obtain the endpoints in the image to be processed.

[0036] In this embodiment, the points on the vascular centerline are traversed. For any point on the vascular centerline, endpoint detection is performed. If the endpoint detection result is an endpoint, the point is marked as an endpoint, and the coordinates of the point are recorded; until all points on the vascular centerline are traversed, all endpoints of the image to be processed are obtained.

[0037] Based on the above embodiment, optionally, the method for endpoint detection includes: judging whether the number of pixel points within the preset three-dimensional neighborhood range of the current point meets the endpoint determination condition. If the endpoint determination condition is met, the current point is determined as an endpoint.

[0038] Among them, the current point refers to the point currently traversed. The preset three-dimensional neighborhood range refers to the neighborhood range in three-dimensional space centered on a certain point. Specifically, the neighborhood range is set by those skilled in the art according to requirements and experience, and is not limited here. In this embodiment, the current neighborhood can be determined according to the position of the current point and the preset three-dimensional neighborhood range, the number of pixel points in the current neighborhood can be determined according to the current neighborhood, and then it can be judged whether the number of pixel points in the current neighborhood all meet the endpoint determination condition. If the endpoint determination condition is met, the current point is determined as an endpoint. Among them, the current neighborhood is the neighborhood in three-dimensional space centered on the current point.

[0039] Exemplarily, assume that the preset three-dimensional neighborhood range is a 3×3×3 neighborhood range. Traverse each point on the blood vessel center line. For any point, determine the number of pixel points within the 3×3×3 neighborhood range centered on the current point. If the number of pixel points is 2, then the current point is an endpoint.

[0040] S130. On the path between any two of the endpoints, determine an anchor point, and perform a local transformation on all points on the path based on the anchor point.

[0041] In this embodiment, randomly select any two endpoints among the endpoints of the image to be detected. On the blood vessel path between the two endpoints, randomly select an anchor point, and perform a local transformation on the points on the blood vessel path based on the anchor point.

[0042] Based on the above embodiment, optionally, the local transformation includes rotation, translation, and scaling; performing the local transformation on all points on the path based on the anchor point includes: forming a target point set based on all points on the path; traversing the target point set, and for any point in the target point set, performing a local transformation based on the pre-transformation coordinates, the coordinates of the anchor point, and the rotation matrix, scaling matrix, and translation vector relative to the anchor point to obtain the post-transformation coordinates.

[0043] Among them, the target point set is the set of all points on the blood vessel path between two endpoints. In this embodiment, a target point set can be formed according to all points on the path between the endpoints, traverse all points in the target point set, and for each traversed point, perform a local transformation on the point according to the pre-transformation coordinates of the point, the coordinates of the anchor point, and the rotation matrix, scaling matrix, and translation vector of the point relative to the anchor point to obtain the post-transformation coordinates of the point. After traversing all points in the target point set, the coordinates of all points on the blood vessel path after local transformation can be obtained, and thus the blood vessel path after local transformation can be determined based on the coordinates of all points on the blood vessel path after local transformation.

[0044] Exemplarily, the local transformation formula is as follows:

[0045]

[0046] R ← RotationMatrix(θ x , θ y , θ z )

[0047] S ← diag(s x , s y , s z )

[0048] B ← (b x , b y , b z )

[0049] Wherein, P hi is the coordinate of the i-th point in the target point set after transformation, p i is the coordinate of the i-th point in the target point set before transformation, p A is the anchor point, set(P sp ) is the target point set, R, S, and B are the rotation matrix, scaling matrix, and translation vector relative to the anchor point p A respectively, and all follow a uniform distribution. θ represents the rotation angle, s represents the scaling factor, b represents the translation amount, ρ r , ρ s and ρ b are the ranges of θ, s, and b respectively, which are set by those skilled in the art according to requirements and are not limited here.

[0050] Based on the above embodiments, optionally, the method further includes: for any point in the target point set, determining the target coordinate based on the transformed coordinate and the local transformation weight.

[0051] Since this application performs local transformation on the path between endpoints rather than global transformation, therefore, for points far from the anchor point, no transformation or little transformation is required, and for points close to the anchor point, transformation is performed. So, this application can use a Gaussian kernel with Euclidean distance as the local transformation weight. In this embodiment, for any point traversed in the target point set, the target coordinate can be determined according to the transformed coordinate and the local transformation weight; wherein, the target coordinate is the transformed coordinate after adding the local transformation weight.

[0052] It can be understood that if the target point is close to the anchor point, then a larger weight can be obtained, and thus the target point can be transformed more; if the target point is far enough from the anchor point, then the obtained weight is approximately 0, and thus the target point does not participate in the transformation, realizing non-rigid transformation for data augmentation. Figure 2 is a schematic diagram of local rotation based on an anchor point provided in Embodiment 1 of the present invention, as Figure 2As shown Figure 2 Taking local rotation based on an anchor point as an example, the points closer to the anchor point in the blood vessel have a larger rotational deformation, while the points farther away from the anchor point have a smaller rotational deformation. Based on the above principle, local rotation is completed.

[0053] Exemplarily, the calculation formula for the target coordinates is as follows:

[0054] P mi = w i * P hi

[0055] The formula for the local transformation weight is as follows:

[0056]

[0057]

[0058] Where P mi represents the target coordinate of the i-th point in the target point set, w i represents the local transformation weight of the i-th point in the target point set, p i represents the coordinate of the i-th point in the target point set before transformation, p A represents the anchor point selected on the blood vessel path, h is the standard deviation of the Gaussian kernel, and π i is a 3×3 diagonal matrix, and z, y, and x on the diagonal all follow the Bernoulli distribution. z, y, and x represent the weights in the corresponding directions respectively.

[0059] It can be understood that since each data augmentation can only perform local transformation on the points on one path, the above operation can be repeated to perform local transformation on multiple paths in the image to be processed, so as to obtain multiple local transformations, thereby ensuring the diversity of data augmentation.

[0060] The technical solution of this embodiment obtains the image to be processed, extracts the blood vessel centerline in the image to be processed; determines the endpoints in the image to be processed based on the blood vessel centerline; on the path between any two endpoints, determines an anchor point, and performs local transformation on all points on the path based on the anchor point. By performing local transformation on the blood vessel path between any two endpoints in the image to be processed based on the anchor point, data augmentation of the image to be processed is realized, the problem of poor training data quality and insufficient training data is solved, the quality of training data is improved, and thus the generalization ability of the model is enhanced.

[0061] Figure 3It is a flowchart of a data enhancement method provided in the second embodiment of the present invention. This embodiment is a preferred embodiment provided on the basis of the above embodiments. Specifically, after determining the endpoints in the image to be processed based on the blood vessel centerline, the method further includes: taking any endpoint in the image to be processed as the first endpoint, and determining the second endpoint closest to the first endpoint based on the first endpoint; on the path between the first endpoint and the second endpoint, determining an anchor point, and performing local transformation on all points on the path based on the anchor point. Among them, the explanations of the same or corresponding terms in the above embodiments are not repeated here. As Figure 3 shown, the method includes:

[0062] S310. Obtain the image to be processed, and extract the blood vessel centerline in the image to be processed.

[0063] S320. Determine the endpoints in the image to be processed based on the blood vessel centerline.

[0064] S330. Take any endpoint in the image to be processed as the first endpoint, and determine the second endpoint closest to the first endpoint based on the first endpoint;

[0065] Among them, the first endpoint is a randomly selected endpoint among all the endpoints in the image to be processed. In this embodiment, a random endpoint is selected from the endpoints in the image to be processed as the first endpoint, and the shortest path is retrieved starting from the first endpoint to obtain the endpoint closest to the first endpoint as the second endpoint. Among them, the retrieval algorithms for the shortest path include but are not limited to the breadth-first search algorithm, the floyd algorithm, etc., which are not limited here.

[0066] It should be noted that when the distance between two endpoints is too far, the topological structure information between the endpoints is easily destroyed. Therefore, when performing local transformation, the blood vessel path between the two closest endpoints is selected for local transformation.

[0067] S340. On the path between the first endpoint and the second endpoint, determine an anchor point, and perform local transformation on all points on the path based on the anchor point.

[0068] In this embodiment, the blood vessel path between the first endpoint and the second endpoint can be determined according to the first endpoint and the second endpoint, a random anchor point is selected on the blood vessel path, and then local transformation is performed on all points on the blood vessel path based on the anchor point.

[0069] In the technical solution of this embodiment, a to-be-processed image is obtained, and the blood vessel centerline in the to-be-processed image is extracted. Endpoints in the to-be-processed image are determined based on the blood vessel centerline. Any endpoint in the to-be-processed image is used as the first endpoint, and a second endpoint closest to the first endpoint is determined based on the first endpoint; an anchor point is determined on the path between the first endpoint and the second endpoint, and all points on the path are locally transformed based on the anchor point. By locally transforming the blood vessel path between the two closest endpoints in the to-be-processed image, data augmentation of the to-be-processed image is realized, the problem of poor quality and insufficient training data is solved, the quality of the training data is improved, and thus the generalization ability of the model is enhanced.

[0070] Figure 4 FIG. 4 is a schematic structural diagram of a data augmentation device provided in Embodiment 3 of the present invention. As Figure 4 shown, the device includes:

[0071] A centerline extraction module 410, configured to obtain a to-be-processed image and extract the blood vessel centerline in the to-be-processed image;

[0072] An endpoint determination module 420, configured to determine endpoints in the to-be-processed image based on the blood vessel centerline;

[0073] A local transformation module 430, configured to determine an anchor point on the path between any two endpoints, and locally transform all points on the path based on the anchor point.

[0074] Based on the above embodiment, optionally, the centerline extraction module 410 is specifically configured to extract the blood vessel skeleton of the to-be-processed image based on a skeleton extraction algorithm to obtain a skeleton image; and determine the blood vessel centerline in the to-be-processed image based on the skeleton image.

[0075] Based on the above embodiment, optionally, the endpoint determination module 420 is specifically configured to traverse the points on the blood vessel centerline, perform endpoint detection on the points on the blood vessel centerline, and obtain the endpoints in the to-be-processed image.

[0076] Based on the above embodiment, optionally, the endpoint determination module 420 includes an endpoint detection unit, configured to determine whether the number of pixel points within a preset three-dimensional neighborhood range of the current point meets an endpoint determination condition. If the endpoint determination condition is met, the current point is determined as an endpoint.

[0077] Based on the above embodiments, optionally, the local transformation includes rotation, translation, and scaling; the local transformation module 430 is specifically configured to form a target point set based on all the points on the path; traverse the target point set, and for any point in the target point set, perform local transformation based on the pre-transformation coordinates, the coordinates of the anchor point, and the rotation matrix, scaling matrix, and translation vector relative to the anchor point to obtain the post-transformation coordinates.

[0078] Based on the above embodiments, optionally, the device further includes a target coordinate determination module, configured to determine a target coordinate for any point in the target point set based on the post-transformation coordinates and the local transformation weight.

[0079] Based on the above embodiments, optionally, the device further includes a second endpoint determination module, configured to use any endpoint in the image to be processed as the first endpoint, and determine a second endpoint closest to the first endpoint based on the first endpoint; the local transformation module 430 is further configured to determine an anchor point on the path between the first endpoint and the second endpoint, and perform local transformation on all the points on the path based on the anchor point.

[0080] The data enhancement device provided by the embodiments of the present invention can execute the data enhancement method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0081] Figure 5 It is a schematic structural diagram of an electronic device provided by Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0082] Such as Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. Among them, the memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0083] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disc, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0084] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the data augmentation method.

[0085] In some embodiments, the data augmentation method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the data augmentation method described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the data augmentation method in any other appropriate way (e.g., by means of firmware).

[0086] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0087] The computer program for implementing the data enhancement method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer programs are executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer programs can be executed entirely on the machine, partially on the machine, executed partially on the machine as an independent software package and partially on a remote machine, or executed entirely on a remote machine or server.

[0088] Embodiment 5 of the present invention also provides a computer-readable storage medium storing computer instructions for causing a processor to execute a data enhancement method, the method including:

[0089] Obtain an image to be processed, and extract the centerline of blood vessels in the image to be processed; determine the endpoints in the image to be processed based on the centerline of blood vessels; on the path between any two endpoints, determine an anchor point, and perform a local transformation on all points on the path based on the anchor point.

[0090] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0091] To provide for interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0092] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0093] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The relationship between the client and the server is created by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0094] It should be understood that various forms of processes shown above can be used, steps can be reordered, added or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0095] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A data augmentation method, characterized in that, including: obtaining an image to be processed, and extracting the centerline of blood vessels in the image to be processed; determining endpoints in the image to be processed based on the centerline of blood vessels; determining an anchor point on a path between any two of the endpoints, and performing local transformation on all points on the path based on the anchor point; after determining the endpoints in the image to be processed based on the centerline of blood vessels, the method further includes: using any endpoint in the image to be processed as a first endpoint, and determining a second endpoint closest to the first endpoint based on the first endpoint; determining an anchor point on a path between the first endpoint and the second endpoint, and performing local transformation on all points on the path based on the anchor point.

2. The method according to claim 1, wherein The extracting the centerline of blood vessels in the image to be processed includes: extracting the blood vessel skeleton of the image to be processed based on a skeleton extraction algorithm to obtain a skeleton image; determining the centerline of blood vessels in the image to be processed based on the skeleton image.

3. The method according to claim 1, wherein The determining the endpoints in the image to be processed based on the centerline of blood vessels includes: traversing the points on the centerline of blood vessels, and performing endpoint detection on the points on the centerline of blood vessels to obtain the endpoints in the image to be processed.

4. The method according to claim 3, wherein The endpoint detection method includes: judging whether the number of pixel points within a preset three-dimensional neighborhood range of the current point meets an endpoint determination condition, and if the endpoint determination condition is met, determining the current point as an endpoint.

5. The method according to claim 1, characterized in that, The local transformation includes rotation, translation, and scaling; the performing local transformation on all points on the path based on the anchor point includes: forming a target point set based on all points on the path; traversing the target point set, and for any point in the target point set, performing local transformation based on the pre-transformation coordinates, the coordinates of the anchor point, and the rotation matrix, scaling matrix, and translation vector relative to the anchor point to obtain post-transformation coordinates.

6. The method according to claim 5, wherein The method further includes: for any point in the target point set, determining target coordinates based on the post-transformation coordinates and local transformation weights.

7. A data enhancement device, characterized in that, including: a centerline extraction module, configured to obtain an image to be processed and extract the centerline of blood vessels in the image to be processed; an endpoint determination module, configured to determine endpoints in the image to be processed based on the centerline of blood vessels; a local transformation module, configured to determine an anchor point on a path between any two of the endpoints and perform local transformation on all points on the path based on the anchor point; wherein, the data augmentation device further includes a second endpoint determination module, configured to use any endpoint in the image to be processed as a first endpoint and determine a second endpoint closest to the first endpoint based on the first endpoint; the local transformation module is further configured to determine an anchor point on a path between the first endpoint and the second endpoint and perform local transformation on all points on the path based on the anchor point.

8. An electronic device, characterized in that, The electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, enables the at least one processor to execute the data enhancement method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the data enhancement method according to any one of claims 1-6.

Images