Arteriovenous point determination method, apparatus, electronic device, and storage medium

Abstract

This disclosure relates to the field of image processing, and more particularly to a method, apparatus, electronic device, and storage medium for determining arteriovenous points. The method includes: segmenting a brain perfusion image sequence into thalamic regions to determine the thalamic regions within the brain perfusion image sequence; determining the signal-time curve corresponding to each candidate arteriovenous point based on the pixel values ​​of the pixel set corresponding to each candidate arteriovenous point in the brain perfusion image; and determining arterial or venous points from the candidate arteriovenous points based on the signal-time curves and, in conjunction with the signal-time curve characteristics of sampled pixels in the thalamic region. This method efficiently and accurately selects healthy arteriovenous points and obtains reasonable AIF and VOF curves, improving analytical accuracy and enhancing the rationality and interpretability of the results.

Description

Technical Field

[0001] This disclosure relates to the field of image processing, and more particularly to a method, apparatus, electronic device, and storage medium for determining arteriovenous points. Background Technology

[0002] Clinically, brain perfusion imaging is mainly used to assess cerebral ischemia and collateral circulation. During this research, a set of healthy arteriovenous points needs to be selected as a reference for subsequent precise calculations. Current methods for selecting arteries and veins primarily rely on manual judgment by physicians or automated search algorithms. Both methods heavily depend on the physician's experience and judgment, making it difficult to determine the optimal result for the region and failing to provide strong evidence to justify the selection of key points. Summary of the Invention

[0003] In view of this, the present disclosure provides a method, apparatus, electronic device, and storage medium for determining arterial and venous points.

[0004] According to one aspect of this disclosure, a method for determining arteriovenous points is provided, comprising: segmenting a brain perfusion image sequence into thalamic regions to determine thalamic regions in the brain perfusion image sequence; determining a signal-time curve corresponding to each arteriovenous candidate point based on the pixel values ​​of the set of pixels corresponding to each arteriovenous candidate point in the brain perfusion image; and determining an arterial point or a venous point from the arteriovenous candidate points based on the signal-time curve and in combination with the signal-time curve features of sampled pixels in the thalamic region.

[0005] In one possible implementation, determining the signal-time curve corresponding to each arteriovenous candidate point based on the pixel values ​​of the set of pixels corresponding to each candidate arteriovenous point in the brain perfusion image includes:

[0006] For each arteriovenous candidate point, in the brain perfusion image, a set of pixels containing a predetermined number of pixels of the arteriovenous candidate point is determined as the set of pixels corresponding to the arteriovenous candidate point.

[0007] Calculate the signal value corresponding to the arterial and vein candidate point based on the pixel values ​​in the pixel set;

[0008] Based on the signal values ​​corresponding to the candidate arteriovenous point in each image of the brain perfusion imaging sequence, the signal-time curve corresponding to the candidate arteriovenous point is determined.

[0009] In one possible implementation, determining an arterial or venous point from the candidate arterial and venous points based on the signal-time curve and in conjunction with the signal-time curve features of sampled pixels in the thalamic region includes:

[0010] Based on preset conditions, the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point is selected from each signal-time curve.

[0011] The candidate arterial or venous points corresponding to the selected signal-time curves are used as arterial or venous points.

[0012] The preset conditions include the relationship between the peak time of the signal-time curve and the average peak time of the signal-time curve of the sampled pixel.

[0013] In one possible implementation, for arterial points, the preset condition is related to one or more of the curve peak value, curve peak time, curve mean, and curve half-width and height, wherein: the curve peak time is not later than the average peak time of the signal-time curve of the sampled pixel point;

[0014] For vein points, the preset conditions are related to one or more of the curve peak value, curve initial value, curve mean value, and curve half-width and height, wherein the curve peak time is not earlier than the average peak time of the signal-time curve of the sampled pixel.

[0015] In one possible implementation, the method further includes:

[0016] Determine the signal-time curves of sampled pixels within the thalamic region;

[0017] Determine the average peak time of the signal-time curve for each sampled pixel.

[0018] In one possible implementation, the thalamic region segmentation of the brain perfusion image sequence includes: segmenting the brain perfusion image sequence into thalamic regions using a thalamic region segmentation model.

[0019] The thalamic region model is based on the Unet network structure and includes an encoder and a decoder. Each encoding module in the encoder includes a residual channel attention module, a ReLU activation function, and a max-pooling downsampling module. Each decoding module in the decoder includes an upsampling module, a residual channel attention module, and a ReLU activation function.

[0020] The residual channel attention module internally uses residual connections to perform residual analysis on the data, and simultaneously uses channel attention to perform weighted calculations on the channel dimensions.

[0021] In one possible implementation, the output data of the ReLU activation function in each encoding module is concatenated with the output data of the upsampling module in the corresponding decoding module, and used as the input of the residual channel attention module of that decoding module.

[0022] According to another aspect of this disclosure, an arteriovenous point determination device is provided, comprising:

[0023] The segmentation module is used to segment the thalamic regions in brain perfusion image sequences and identify the thalamic regions in the brain perfusion image sequences.

[0024] The curve determination module is used to determine the signal-time curve corresponding to each arteriovenous candidate point based on the pixel value of the set of pixels corresponding to each arteriovenous candidate point in the brain perfusion image.

[0025] The arteriovenous point determination module is used to determine arterial or venous points from the candidate arteriovenous points based on the signal-time curve and the signal-time curve features of sampled pixels in the thalamic region.

[0026] According to another aspect of this disclosure, an electronic device is provided, characterized in that it comprises:

[0027] processor;

[0028] Memory used to store processor-executable instructions;

[0029] The processor is configured to implement the above method when executing instructions stored in the memory.

[0030] According to another aspect of this disclosure, a non-volatile computer-readable storage medium is provided, on which computer program instructions are stored, characterized in that the computer program instructions implement the above-described method when executed by a processor.

[0031] By identifying the thalamic region in the brain perfusion imaging sequence and determining the signal-time curves corresponding to each arteriovenous candidate point based on the corresponding pixel values ​​in the brain perfusion images, the determination of the time signal curves considers not only the arteriovenous candidate points themselves but also the set of pixels containing the arteriovenous candidate points. Based on the signal-time curves and combined with the signal-time curve characteristics of the sampled pixels in the thalamic region, arterial or venous points are determined from the arteriovenous candidate points, thereby efficiently and accurately selecting healthy arteriovenous points and obtaining reasonable AIF and VOF curves, improving the accuracy of analysis and enhancing the rationality and interpretability of the results.

[0032] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0033] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0034] Figure 1A flowchart of a method for determining arteriovenous points according to an embodiment of this application is shown.

[0035] Figure 2 A schematic diagram of a thalamic region segmentation model according to an embodiment of this application is shown.

[0036] Figure 3 A schematic diagram of a residual channel attention module according to an embodiment of this application is shown.

[0037] Figure 4 This diagram illustrates the structure of an arteriovenous point determination device according to an embodiment of this application.

[0038] Figure 5 This is a block diagram illustrating an apparatus 1900 for determining arteriovenous points according to an exemplary embodiment. Detailed Implementation

[0039] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0040] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0041] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0042] Due to its clinical value in assessing cerebral ischemia and collateral circulation, brain perfusion imaging has received widespread attention and application. Brain perfusion imaging includes CT perfusion imaging (CTP) and perfusion-weighted imaging (PWI). During the research process, it is often necessary to select a set of healthy arteriovenous points (one arterial point and one venous point) as a reference for subsequent precise calculations. The arterial input function (AIF) and venous output function (VOF) curves corresponding to the selected vascular points have significant direct implications for the rationality and accuracy of the final results.

[0043] Current methods for selecting arteriovenous points primarily rely on doctors' manual judgment to select healthy points. A few algorithms and software also use prior knowledge or model output to determine candidate regions and then search within them. A cost function is constructed based on the correlation features of the contrast agent concentration-time curves corresponding to each pixel to select the final arteriovenous point. Because the regions of interest for arteriovenous points are large, and image annotation tools cannot directly observe the contrast agent concentration-time curves of the selected vessels, repeated comparisons and selections of points with good AIF and VOF curves are necessary in practice, resulting in significant waste of manual labor. Furthermore, due to the cumbersome comparison process, doctors often only select a few vessels with obvious features for comparison, so the final result is often not the optimal result for that region. Existing automatic search algorithms mostly rely on prior regions or model outputs to determine candidate regions, and then construct a cost function based on the one-dimensional features of the contrast agent concentration-time curves corresponding to each pixel within the candidate region to evaluate the curves. While this method can achieve automatic selection, the manually designed cost function becomes insufficient when dealing with large amounts of real-world data. Most existing automatic selection methods consider the individual pixel itself, without taking into account the relationship between arterial and venous points and their corresponding curves, and cannot provide strong evidence to prove the rationality of the selection of key points.

[0044] This application proposes a method, apparatus, electronic device, and storage medium for determining arteriovenous points. It determines the thalamic region in a brain perfusion image sequence and the signal-time curve corresponding to each candidate arteriovenous point based on the pixel values ​​of the brain perfusion image. When determining the time-signal curve, it considers not only the candidate arteriovenous point itself but also the set of pixels containing the candidate arteriovenous point. Based on the signal-time curve and combined with the signal-time curve characteristics of the sampled pixels in the thalamic region, it determines arterial or venous points from the candidate arteriovenous points. This efficiently and accurately selects healthy arteriovenous points and obtains reasonable AIF and VOF curves, improving analysis accuracy and enhancing the rationality and interpretability of the results.

[0045] Figure 1 A flowchart of a method for determining arteriovenous points according to an embodiment of this application is shown, as follows: Figure 1 As shown, the method includes:

[0046] S100, thalamic region segmentation is performed on the brain perfusion image sequence to identify the thalamic regions in the brain perfusion image sequence.

[0047] S200, determine the signal-time curve corresponding to each arterial and venous candidate point based on the pixel value of the pixel set corresponding to each arterial and venous candidate point in the brain perfusion image;

[0048] S300, based on the signal-time curve and combined with the signal-time curve features of the sampled pixels in the thalamic region, determine the arterial or venous point from the candidate arterial and venous points.

[0049] Brain perfusion imaging sequence refers to an image sequence obtained through brain perfusion imaging technology. Brain perfusion imaging utilizes drug injection into the body, and imaging instruments are used to obtain images reflecting brain function and metabolism. A brain perfusion imaging sequence may include a sequence of multiple images corresponding to different imaging times, and each pixel in the sequence may correspond to a contrast agent concentration-time curve. For example, based on brain perfusion imaging technology, continuous dynamic scanning of a selected area of ​​interest can be performed to obtain the contrast agent concentration-time curve for each pixel of the selected area, and hemodynamic parameters and perfusion image representation can be obtained through mathematical modeling.

[0050] Brain perfusion imaging sequences may include, for example, CTP sequences based on CTP technology and PWI sequences based on PWI technology. CTP involves performing continuous CT scans on the region of interest (ROI) during rapid intravenous perfusion of contrast agent to obtain the contrast agent concentration-time curve for the RIO, and then using different mathematical models to calculate various perfusion parameter values. PWI is a type of magnetic resonance imaging that uses intravenous injection of paramagnetic contrast agent to reflect tissue blood perfusion function by calculating perfusion parameters, thus providing functional imaging that reflects changes in tissue hemodynamics.

[0051] In step S100, the brain perfusion image sequence can be segmented into thalamic regions using a thalamic region segmentation model to determine the thalamic regions in the brain perfusion image sequence.

[0052] Figure 2 A schematic diagram of a thalamic region segmentation model according to an embodiment of this application is shown, as follows: Figure 2 As shown, the thalamic region model is based on the Unet network structure and includes an encoder and a decoder. Each encoding module in the encoder includes a residual channel attention module, a ReLU activation function, and a max pooling downsampling module. Each decoding module in the decoder includes an upsampling module, a residual channel attention module, and a ReLU activation function. The residual channel attention module uses residual connections to perform residual analysis on the data and uses channel attention to perform weighted calculations on the channel dimensions.

[0053] By employing the above methods, the accuracy of thalamic region selection can be improved, thereby enhancing the interpretability of the model.

[0054] See Figure 2In one example, each encoding module in the encoder contains two residual channel attention modules (Res_SE), a ReLU activation function (ReLU), and one max-pooling downsampling operation, where the number of feature map channels doubles after downsampling. Each decoding module in the decoder contains sequential upsampling (deconvolution), two residual channel attention modules (Res_SE), and a ReLU activation function (ReLU), where the number of feature map channels is halved after each upsampling. The feature map before downsampling is concatenated with the decoder part. During the transition from encoder to decoder, two residual channel attention modules and one ReLU activation function are applied. Finally, the decoder output is dimensionality-reduced using a 1×1 convolution, and the Softmax function is used to output the probability of each pixel for each class; the class with the highest probability is the corresponding segmentation result.

[0055] In one possible implementation, the output data of the ReLU activation function in each encoding module is concatenated with the output data of the upsampling module in the corresponding decoding module, and used as the input to the residual channel attention module of that decoding module. For example Figure 2 As shown, the output of the encoder Res_SE*2+ReLu is concatenated with the output of the decoder Upsample, and used as the input of the decoder Res_SE*2+ReLu.

[0056] Figure 3 A schematic diagram of a residual channel attention module according to an embodiment of this application is shown. The gradient vanishing problem is greatly alleviated by using residual connections, while channel attention is used to perform weighted calculations on the channel dimensions (see [link]). Figure 3 In this process, channels consisting of global pooling, fully connected (FC), and sigmoid activation functions are used to improve the network's feature extraction capabilities and training speed.

[0057] In one possible implementation, step S200 involves determining the signal-time curve corresponding to each arteriovenous candidate point based on the pixel values ​​of the pixel set corresponding to each arteriovenous candidate point in the brain perfusion image, including:

[0058] For each arteriovenous candidate point, in the brain perfusion image, a set of pixels containing a predetermined number of pixels of the arteriovenous candidate point is determined as the set of pixels corresponding to the arteriovenous candidate point.

[0059] Calculate the signal value corresponding to the arterial and vein candidate point based on the pixel values ​​in the pixel set;

[0060] Based on the signal values ​​corresponding to the candidate arteriovenous point in each image of the brain perfusion imaging sequence, the signal-time curve corresponding to the candidate arteriovenous point is determined.

[0061] Artery and vein candidate points can include artery candidate points and vein candidate points, which can be specified manually or obtained through methods such as the following. First, the candidate centers of artery and vein points can be determined. These candidate centers can be specified manually or extracted from image data. Then, samples are taken at intervals of three pixels (adjustable according to actual conditions) within a distance of 20 pixels around each candidate center of artery and vein point to obtain a set of candidate artery and vein points.

[0062] For each arteriovenous candidate point, the pixels in its eight neighboring regions can be selected from each image in the brain perfusion imaging sequence as the aforementioned pixel set. The selection method of the pixel set can be adjusted according to the actual situation. The mean value of the eight neighboring pixels can be used to replace the pixel value of the arteriovenous candidate point itself to plot the signal-time curve of the arteriovenous candidate point.

[0063] This approach fully considers the surrounding conditions of the artery candidate site, thus improving accuracy.

[0064] In one possible implementation, step S300, determining an arterial or venous point from the candidate arterial and venous points based on the signal-time curve and in conjunction with the signal-time curve features of the sampled pixels in the thalamic region, includes:

[0065] Based on preset conditions, signal-time curves corresponding to arterial points or venous points are selected from each signal-time curve; the candidate arterial or venous points corresponding to the selected signal-time curves are then used as arterial or venous points. The preset conditions include the relationship between the peak time of the signal-time curve and the average peak time of the signal-time curve of the sampled pixel.

[0066] In one possible implementation, for arterial points, the preset condition is related to one or more of the curve peak value, curve peak time, curve mean, and curve half-width and height, wherein the curve peak time is not later than the average peak time of the signal-time curve of the sampled pixel; for venous points, the preset condition is related to one or more of the curve peak value, curve initial value, curve mean, and curve half-width and height, wherein the curve peak time is not earlier than the average peak time of the signal-time curve of the sampled pixel.

[0067] For example, for vein candidate points, based on prior information about the actual blood vessel line shape, a set of reasonable signal-time curves can be selected by limiting parameters such as curve peak value, initial value, mean value, and half-width and height. That is, the peak value, initial value, mean value, and half-width and height of the selected signal-time curves are within a preset range. This preset range can be reasonably set based on the above prior information. At the same time, it is required that the curve peak time is not earlier than the average peak time of the sampled pixels. Among multiple curves that meet the above conditions, the curve with the highest peak value can be selected as the VOF curve and the corresponding vein point (arteriovenous candidate point) can be output as the final vein point.

[0068] For artery candidate points, based on prior information about the actual vessel point shape, the peak value and peak time of VOF can be used as constraints. Simultaneously, combined with its mean and half-width / height, a set of reasonable signal-time curves can be selected. Specifically, the peak value, peak time, mean, and half-width / height of the selected signal-time curves should fall within a preset range, where this preset range can be reasonably set based on the aforementioned prior information. Furthermore, the peak time of the curve must not be later than the average peak time of the sampled pixels. Among multiple curves meeting these conditions, the curve with the highest peak value is ultimately selected as the AIF curve, and the corresponding artery point (arteriovenous candidate point) is output as the final artery point.

[0069] In one possible implementation, the average peak time of the sampled pixels can be obtained in the following way, that is, the method described above further includes:

[0070] Determine the signal-time curves of sampled pixels within the thalamic region;

[0071] Determine the average peak time of the signal-time curve for each sampled pixel.

[0072] In one example, samples can be taken at intervals of three pixels (which can be adjusted according to the actual situation) within the thalamic region to obtain sampled pixels. The average value of the eight neighboring pixels is then used to replace the original pixel value (if the modality is PWI, the first three phases need to be used as the base value, and the pixel value is processed by logarithmic transformation to make the pixel value proportional to the contrast agent concentration). Contrast agent concentration-time curves (signal-time curves) are plotted at each point. The average peak time is obtained based on the curves of all sampled pixels and used as the dividing point of the arteriovenous phase of the curve.

[0073] The following is an application example of Embodiment 1 of this application.

[0074] First, the thalamic segmentation model can be trained. Acquire a certain amount of initial brain perfusion imaging data (data from the first scan, before the contrast agent reaches the cerebral blood vessels), and label the thalamic regions, for example, assigning 1 to the thalamus and 0 to other regions, thus obtaining a sample dataset. This can be trained using CTP sequences to obtain a thalamic segmentation model for CTP sequences, or using PWI sequences to obtain a thalamic segmentation model for PWI sequences. Specific training steps may include:

[0075] (1) Randomly batch-sample several data points from the sample dataset and take a 2D slice along the z-axis as the original data.

[0076] (2) Perform data preprocessing and data normalization on the sample data.

[0077] (3) Perform data augmentation on the preprocessed and normalized sample data, such as random rotation, random scaling, and elastic deformation.

[0078] (4) Input the augmented sample data into the network model to obtain the thalamic segmentation mask.

[0079] (5) Calculate the loss using the weighted sum of Dice loss and cross-entropy loss as the loss function, and perform gradient calculation, backpropagation, and update the network model parameters.

[0080] (6) Repeat (1)-(5) until the model parameters converge to obtain the thalamus segmentation model.

[0081] Then, predictions are made using a thalamic segmentation model, and the specific steps may include:

[0082] (1) Input the first phase data of the brain perfusion imaging sequence (which may be the part of the first phase data after removing the above sample data).

[0083] (2) Data preprocessing and normalization were performed on the 2D slices of the first phase data of the brain perfusion imaging sequence.

[0084] (3) Input the preprocessed and normalized brain perfusion image sequence into the thalamus segmentation model, set the confidence level to 0.5, and obtain the final thalamus segmentation result based on the confidence level, that is, the thalamus region in each frame of the brain perfusion image sequence.

[0085] (1) Output thalamic segmentation mask (foreground is 1, background is 0)

[0086] Sampled pixels were selected in the segmented thalamic region, and the average peak time of the sampled pixels was obtained in the manner described above.

[0087] Based on the artificially designated candidate centers of arteriovenous points, candidate arterial points are determined in the manner described above. The signal-time curves corresponding to each candidate arteriovenous point are determined. Based on the preset screening conditions described above, the signal-time curves are selected, and the corresponding arterial or venous points are determined.

[0088] Figure 4 This diagram illustrates a structural diagram of an arteriovenous point determination device according to an embodiment of this application, comprising:

[0089] The segmentation module 401 is used to segment the thalamic region in the brain perfusion image sequence and determine the thalamic region in the brain perfusion image sequence.

[0090] The curve determination module 402 is used to determine the signal-time curve corresponding to each arteriovenous candidate point based on the pixel value of the set of pixels corresponding to each arteriovenous candidate point in the brain perfusion image.

[0091] The arteriovenous point determination module 403 determines arterial or venous points from the candidate arteriovenous points based on the signal-time curve and the signal-time curve features of the sampled pixels in the thalamic region.

[0092] In one possible implementation, determining the signal-time curve corresponding to each arteriovenous candidate point based on the pixel values ​​of the set of pixels corresponding to each candidate arteriovenous point in the brain perfusion image includes:

[0093] For each arteriovenous candidate point, in the brain perfusion image, a set of pixels containing a predetermined number of pixels of the arteriovenous candidate point is determined as the set of pixels corresponding to the arteriovenous candidate point.

[0094] Calculate the signal value corresponding to the arterial and vein candidate point based on the pixel values ​​in the pixel set;

[0095] Based on the signal values ​​corresponding to the candidate arteriovenous point in each image of the brain perfusion imaging sequence, the signal-time curve corresponding to the candidate arteriovenous point is determined.

[0096] In one possible implementation, determining the arterial or venous point from the candidate arterial and venous points based on the signal-time curve and in conjunction with the signal-time curve features of sampled pixels in the thalamic region includes:

[0097] Based on preset conditions, the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point is selected from each signal-time curve.

[0098] The candidate arterial or venous points corresponding to the selected signal-time curves are used as arterial or venous points.

[0099] The preset conditions include the relationship between the peak time of the signal-time curve and the average peak time of the signal-time curve of the sampled pixel.

[0100] In one possible implementation, for arterial points, the preset condition is related to one or more of the curve peak value, curve peak time, curve mean, and curve half-width and height, wherein: the curve peak time is not later than the average peak time of the signal-time curve of the sampled pixel point;

[0101] For vein points, the preset conditions are related to one or more of the curve peak value, curve initial value, curve mean value, and curve half-width and height, wherein the curve peak time is not earlier than the average peak time of the signal-time curve of the sampled pixel.

[0102] In one possible implementation, the device further includes:

[0103] The sampling point curve determination module is used to determine the signal-time curve of the sampling pixels in the thalamic region;

[0104] The average peak time determination module is used to determine the average peak time of the signal-time curve for each sampled pixel.

[0105] In one possible implementation, the thalamic region segmentation of the brain perfusion image sequence includes: segmenting the brain perfusion image sequence into thalamic regions using a thalamic region segmentation model.

[0106] The thalamic region model is based on the Unet network structure and includes an encoder and a decoder. Each encoding module in the encoder includes a residual channel attention module, a ReLU activation function, and a max-pooling downsampling module. Each decoding module in the decoder includes an upsampling module, a residual channel attention module, and a ReLU activation function.

[0107] The residual channel attention module internally uses residual connections to perform residual analysis on the data, and simultaneously uses channel attention to perform weighted calculations on the channel dimensions.

[0108] In one possible implementation, the output data of the ReLU activation function in each encoding module is concatenated with the output data of the upsampling module in the corresponding decoding module, and used as the input of the residual channel attention module of that decoding module.

[0109] This application also proposes an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0110] This application also proposes a non-volatile computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, implement the above-described method.

[0111] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0112] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0113] Figure 5 This is a block diagram illustrating an apparatus 1900 for arteriovenous point determination according to an exemplary embodiment. For example, apparatus 1900 may be provided as a server or terminal device. Referring to the figures, apparatus 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions executable by the processing component 1922, such as application programs. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, processing component 1922 is configured to execute instructions to perform the methods described above.

[0114] Device 1900 may also include a power supply component 1926 configured to perform power management of device 1900, a wired or wireless network interface 1950 configured to connect device 1900 to a network, and an input / output interface 1958 (I / O interface). Device 1900 can operate on an operating system stored in memory 1932, such as Windows Server™, MacOS X™, Unix™, Linux™, FreeBSD™, or similar.

[0115] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of the device 1900 to perform the above-described method.

[0116] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0117] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0118] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0119] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0120] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0121] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0122] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0124] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for determining arteriovenous points, characterized in that, include: Thalamic regions were segmented from brain perfusion imaging sequences to identify the thalamic regions within the sequences. Based on the pixel values ​​of the pixel set corresponding to each arteriovenous candidate point in the brain perfusion image, the signal-time curve corresponding to each arteriovenous candidate point is determined. Based on the signal-time curve and combined with the signal-time curve features of the sampled pixels in the thalamic region, arterial or venous points are determined from the candidate arterial and venous points. The step of determining arterial or venous points from the candidate arterial and venous points based on the signal-time curve and the signal-time curve features of sampled pixels in the thalamus region includes: Based on preset conditions, the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point is selected from each signal-time curve. The candidate arterial or venous points corresponding to the selected signal-time curves are used as arterial or venous points. The step of selecting the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point from each signal-time curve according to preset conditions includes: For an artery point, based on prior information about the actual blood vessel line shape, a preset range for the artery point is determined. At least one signal-time curve is selected from the signal-time curves. One or more of the peak value, peak time, mean value, and half-width / highest value of each selected signal-time curve are within the preset range for the corresponding artery point, and the peak time is not later than the average peak time of the signal-time curves of the sampled pixels. The signal-time curve with the highest peak value is selected from the at least one signal-time curve as the signal-time curve corresponding to the artery point. For a vein point, based on prior information about the actual blood vessel line shape, a preset range for the vein point is determined. At least one signal-time curve is selected from the signal-time curves. One or more of the peak value, initial value, mean value, and half-width / highest value of each selected signal-time curve are within the preset range for the vein point, and the peak time is not earlier than the average peak time of the signal-time curves of the sampled pixels. The signal-time curve with the highest peak value is selected from the at least one signal-time curve as the signal-time curve corresponding to the vein point.

2. The method according to claim 1, characterized in that: The step of determining the signal-time curve corresponding to each arteriovenous candidate point based on the pixel values ​​of the pixel set corresponding to each arteriovenous candidate point in the brain perfusion image includes: For each arteriovenous candidate point, in the brain perfusion image, a set of pixels containing a predetermined number of pixels of the arteriovenous candidate point is determined as the set of pixels corresponding to the arteriovenous candidate point. Calculate the signal value corresponding to the arterial and vein candidate point based on the pixel values ​​in the pixel set; Based on the signal values ​​corresponding to the candidate arteriovenous point in each image of the brain perfusion imaging sequence, the signal-time curve corresponding to the candidate arteriovenous point is determined.

3. The method according to claim 1, characterized in that, The method further includes: Determine the signal-time curves of sampled pixels within the thalamic region; Determine the average peak time of the signal-time curve for each sampled pixel.

4. The method according to claim 1, characterized in that, The thalamic region segmentation of the brain perfusion image sequence includes: segmenting the brain perfusion image sequence into thalamic regions using a thalamic region segmentation model; The thalamic region segmentation model is based on the Unet network structure and includes an encoder and a decoder. Each encoding module in the encoder includes a residual channel attention module, a ReLU activation function, and a max-pooling downsampling module. Each decoding module in the decoder includes an upsampling module, a residual channel attention module, and a ReLU activation function. The residual channel attention module internally uses residual connections to perform residual analysis on the data, and simultaneously uses channel attention to perform weighted calculations on the channel dimensions.

5. The method according to claim 4, characterized in that, The output data of the ReLU activation function in each encoding module is concatenated with the output data of the upsampling module in the corresponding decoding module, and used as the input of the residual channel attention module of that decoding module.

6. A device for determining arteriovenous points, characterized in that, include: The segmentation module is used to segment the thalamic regions in brain perfusion image sequences and identify the thalamic regions in the brain perfusion image sequences. The curve determination module determines the signal-time curve corresponding to each arterial and venous candidate point based on the pixel values ​​of the pixel set corresponding to each arterial and venous candidate point in the brain perfusion image. The arteriovenous point determination module determines arterial or venous points from the candidate arteriovenous points based on the signal-time curve and the signal-time curve features of sampled pixels in the thalamic region. The step of determining arterial or venous points from the candidate arterial and venous points based on the signal-time curve and the signal-time curve features of sampled pixels in the thalamus region includes: Based on preset conditions, the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point is selected from each signal-time curve. The candidate arterial or venous points corresponding to the selected signal-time curves are used as arterial or venous points. The step of selecting the signal-time curve corresponding to the arterial point or the signal-time curve corresponding to the venous point from each signal-time curve according to preset conditions includes: For an artery point, based on prior information about the actual blood vessel line shape, a preset range for the artery point is determined. At least one signal-time curve is selected from the signal-time curves. One or more of the peak value, peak time, mean value, and half-width / highest value of each selected signal-time curve are within the preset range for the corresponding artery point, and the peak time is not later than the average peak time of the signal-time curves of the sampled pixels. The signal-time curve with the highest peak value is selected from the at least one signal-time curve as the signal-time curve corresponding to the artery point. For a vein point, based on prior information about the actual blood vessel line shape, a preset range for the vein point is determined. At least one signal-time curve is selected from the signal-time curves. One or more of the peak value, initial value, mean value, and half-width / highest value of each selected signal-time curve are within the preset range for the vein point, and the peak time is not earlier than the average peak time of the signal-time curves of the sampled pixels. The signal-time curve with the highest peak value is selected from the at least one signal-time curve as the signal-time curve corresponding to the vein point.

7. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to implement the method of any one of claims 1 to 5 when executing instructions stored in the memory.

8. A non-volatile computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 5.

Images