Scanning identification method and device, computer device and storage medium

Abstract

The application relates to a scanning identification method and device, computer equipment and a storage medium. The method comprises the following steps: in response to a scanning instruction, controlling a target device to perform plain scanning and arterial phase scanning and dynamic feature identification respectively, obtaining first scanning data and first dynamic features corresponding to the first scanning data; if the first dynamic features do not match preset internal leakage features, controlling the target device to perform delay phase scanning and dynamic feature identification, obtaining second scanning data and second dynamic features corresponding to the second scanning data; and determining a scanning identification result based on the first dynamic features or the second dynamic features. The method can improve the scanning identification accuracy of subtle features.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and in particular to a scanning and identifying method and device, a computer device and a storage medium. BACKGROUND

[0002] With the development of medical imaging technology, disease identification based on multi-modal scanning has become an indispensable part of clinical practice.

[0003] Traditional scanning methods mainly rely on single-period imaging, which is difficult to fully capture the characteristics of the lesion area, resulting in limited identification accuracy of subtle lesions (such as internal leaks). In order to solve the above problems, the traditional technology usually increases the number of scans or prolongs the single scan time to obtain more data, however, this scheme has the problems of large image registration error, high radiation dose, and low identification efficiency, which makes it difficult to accurately identify the lesion boundary and nature in the dynamic contrast agent distribution change.

[0004] Therefore, there is still a problem of low scanning and identifying accuracy of subtle features in the prior art. SUMMARY

[0005] Therefore, it is necessary to provide a scanning and identifying method, device, computer device and storage medium capable of improving the scanning and identifying accuracy of subtle features in view of the above technical problems.

[0006] In a first aspect, the present application provides a scanning and identifying method, which comprises:

[0007] In response to a scanning instruction, a target device is controlled to perform plain scan and arterial phase scan and dynamic feature identification respectively, to obtain first scanning data and first dynamic features corresponding to the first scanning data;

[0008] If the first dynamic features do not match the preset internal leak features, the target device is controlled to perform delayed phase scan and dynamic feature identification, to obtain second scanning data and second dynamic features corresponding to the second scanning data;

[0009] Based on the first dynamic features or the second dynamic features, a scanning and identifying result is determined.

[0010] In one embodiment, the target device is a multi-slice CT device, a dual-source CT device or a magnetic resonance imaging device.

[0011] In one embodiment, the control of the target device to perform plain scan and arterial phase scan and dynamic feature identification respectively to obtain first scanning data and first dynamic features corresponding to the first scanning data comprises:

[0012] The target device is controlled to perform a plain scan and an arterial phase scan respectively to obtain first scan data; the first scan data includes plain scan data and arterial phase scan data;

[0013] The plain scan data and the arterial phase scan data are registered based on a preset registration algorithm to obtain first registered scan data;

[0014] Dynamic feature extraction is performed based on the first registered scan data to obtain the first dynamic feature.

[0015] In one of the embodiments, the controlling the target device to perform a delay phase scan and dynamic feature recognition to obtain second scan data and a second dynamic feature corresponding to the second scan data further includes:

[0016] The target device is controlled to perform a delay phase scan to obtain second scan data; the second scan data includes delay phase scan data;

[0017] The delay phase scan data and the arterial phase scan data are registered based on a preset registration algorithm to obtain second registered scan data;

[0018] Dynamic feature extraction is performed based on the second registered scan data to obtain the second dynamic feature.

[0019] In one of the embodiments, the dynamic feature extraction based on the second registered scan data to obtain the second dynamic feature includes:

[0020] Based on the second registered scan data, a CT value difference between the delay phase and the arterial phase is calculated to obtain a HU difference value;

[0021] Based on a scan time difference of the second registered scan data, a concentration change per unit time is calculated to obtain an accumulation rate;

[0022] Based on the second registered scan data and a preset voxel window, a gray scale distribution complexity is calculated to obtain a morphological entropy value;

[0023] Based on a Euclidean distance of a voxel to a target component in the second registered scan data, a sensitivity matrix is generated to obtain a spatial weight;

[0024] Based on the HU difference value, the accumulation rate, the morphological entropy value and the spatial weight, a shape feature map is generated;

[0025] Based on the arterial phase scan data and the shape feature map, an input is made to a pre-trained neural network to obtain the second dynamic feature.

[0026] In one of the embodiments, the controlling the target device to perform the delay phase scanning and the dynamic feature identification to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data comprises:

[0027] determining the scanning parameter of the target device based on the scanning instruction and the object information of the target object;

[0028] controlling the target device to perform the delay phase scanning based on the scanning parameter to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data.

[0029] In one of the embodiments, the second scanning data comprises first delay phase scanning data and second delay phase scanning data; and the controlling the target device to perform the delay phase scanning and the dynamic feature identification to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data further comprises:

[0030] controlling the target device to perform the delay phase scanning and the dynamic feature identification to obtain the first delay phase scanning data and the first delay phase dynamic feature corresponding to the first delay phase scanning data;

[0031] if the first delay phase dynamic feature does not match the preset internal leakage feature, controlling the target device to perform the delay phase scanning and the dynamic feature identification to obtain the second delay phase scanning data and the second delay phase dynamic feature corresponding to the second delay phase scanning data;

[0032] determining the second dynamic feature based on the first delay phase dynamic feature and the second delay phase dynamic feature.

[0033] In a second aspect, the present application provides a scanning and identifying device, which comprises:

[0034] a first scanning module, configured to control a target device to perform the plain scanning and the arterial phase scanning and the dynamic feature identification respectively in response to a scanning instruction to obtain the first scanning data and the first dynamic feature corresponding to the first scanning data;

[0035] a second scanning module, configured to control the target device to perform the delay phase scanning and the dynamic feature identification to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data if the first dynamic feature does not match the preset internal leakage feature;

[0036] a result determining module, configured to determine a scanning and identifying result based on the first dynamic feature or the second dynamic feature.

[0037] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method as described above when executing the computer program.

[0038] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method as described above.

[0039] The scanning identification method, device, computer device and storage medium described above, by responding to a scanning instruction, control a target device to perform plain scanning and arterial phase scanning and dynamic feature identification respectively, obtain first scanning data and first dynamic features corresponding to the first scanning data; if the first dynamic features do not match preset internal leakage features, control the target device to perform delay phase scanning and dynamic feature identification, obtain second scanning data and second dynamic features corresponding to the second scanning data; and determine a scanning identification result based on the first dynamic features or the second dynamic features. Through the execution of multi-phase scanning and dynamic feature analysis, the combination of matching judgment of preset internal leakage features and conditional delay scanning strategy, and the use of the complementarity of multi-time image data for feature fusion and classification modeling, the limitations of single arterial phase scanning can be made up, different internal leakage types can be distinguished through dynamic feature comparison and analysis, and radiation exposure and artificial errors can be reduced through image registration and feature fusion, so that the effect of improving the scanning identification accuracy of subtle features is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 An application environment diagram of the scanning identification method in an embodiment;

[0041] Figure 2 A flowchart of the scanning identification method in an embodiment;

[0042] Figure 3 A flowchart of the scanning identification method in another embodiment;

[0043] Figure 4 A schematic diagram of arterial phase scanning data in one view in an embodiment;

[0044] Figure 5 A schematic diagram of delay phase scanning data in one view in an embodiment;

[0045] Figure 6 A schematic diagram of arterial phase scanning data in another view in an embodiment;

[0046] Figure 7 A schematic diagram of delay phase scanning data in another view in an embodiment;

[0047] Figure 8A schematic diagram of a three-dimensional reconstructed image in one embodiment;

[0048] Figure 9 A schematic diagram of another three-dimensional reconstructed image in one embodiment;

[0049] Figure 10 A block diagram of the structure of a scanning identification device in one embodiment;

[0050] Figure 11 An internal structure diagram of a computer device in one embodiment. DETAILED DESCRIPTION

[0051] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0052] The scanning identification method provided by the embodiments of the present application can be applied in an application environment as shown in Figure 1 . The terminal 102 communicates with the server 104 through a network. The data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The terminal 102 controls the target device to perform flat scan and arterial phase scan and dynamic feature identification in response to a scanning instruction, obtains first scanning data and first dynamic features corresponding to the first scanning data; if the first dynamic features do not match preset internal leakage features, controls the target device to perform delay phase scan and dynamic feature identification, obtains second scanning data and second dynamic features corresponding to the second scanning data; and determines a scanning identification result based on the first dynamic features or the second dynamic features. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones and tablet computers. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0053] In one embodiment, as shown in Figure 2 , a scanning identification method is provided, and the scanning identification method comprises:

[0054] Step S100, in response to a scanning instruction, controlling a target device to perform flat scan and arterial phase scan and dynamic feature identification, obtaining first scanning data and first dynamic features corresponding to the first scanning data.

[0055] The scanning instruction can be a control instruction containing general parameters, such as scanning site, contrast agent type, etc., which can be generated by user input or system preset.

[0056] The plain scan can be a conventional CT scan without injection of contrast agent, which is used to obtain basic anatomical information of the target region, and can be obtained through the original tomographic image generated after X-ray penetrates the human body. Illustratively, the plain scan can include visualization of static structural features such as bones, soft tissues, etc. The arterial phase scan is used to capture the morphology and hemodynamic characteristics when the contrast agent concentration in the blood vessel reaches the peak.

[0057] The dynamic feature recognition can be analysis of the arterial phase scan data by an image processing algorithm to extract dynamic feature parameters such as blood vessel wall integrity, blood flow signal intensity, contrast agent distribution pattern, etc. Illustratively, the dynamic feature recognition can be achieved by noise suppression, blood vessel segmentation, blood flow velocity calculation, etc., so as to obtain a set of quantitative feature parameters as the first dynamic feature, which serves as the basis for judging whether there is an internal leakage or other abnormalities.

[0058] In step S200, if the first dynamic feature does not match the preset internal leakage feature, the target device is controlled to perform a delay phase scan and dynamic feature recognition to obtain second scan data and a second dynamic feature corresponding to the second scan data.

[0059] The preset internal leakage feature can be a typical internal leakage feature data determined based on prior knowledge, which can include hemodynamic parameter thresholds of different types of internal leakage in the arterial phase. Illustratively, the preset internal leakage feature can include parameters such as contrast agent retention time and concentration gradient change in a specific region.

[0060] It can be understood that when the first dynamic feature does not match the preset internal leakage feature, it means that the internal leakage feature cannot be determined in the current first scan data. In this embodiment, further judgment is made through the delay phase scan.

[0061] The delay phase scan can be a CT scan performed again within a certain time interval after the arterial phase scan, which can be achieved by adjusting the scan delay time parameter and then performing the scan. Illustratively, the delay phase scan can capture the dynamic distribution change of the contrast agent in the tissue gap.

[0062] In this embodiment, the control of the target device to perform the delay phase scan and the dynamic feature recognition can be to obtain the second scan data through the delay phase scan, and then to perform the dynamic feature recognition on the second scan data to obtain the second dynamic feature. Further, when performing the dynamic feature recognition, the feature parameters related to the delay in the second scan data, such as the continuous accumulation of contrast agent in a specific region and abnormal flow path, can be analyzed to identify the internal leakage type that is not visible in the arterial phase, and the development law of the region of interest can be determined through time series analysis.

[0063] In step S300, a scan recognition result is determined based on the first dynamic feature or the second dynamic feature.

[0064] The determination of the scan recognition result can be comprehensive analysis of multi-phase dynamic characteristics. For example, the arterial phase image and the delayed phase image can be registered and aligned to eliminate deviation, the feature parameters of the two phases are integrated by a feature fusion algorithm, the integrated features are classified and located by a classification model, and thus the scan recognition result is obtained.

[0065] The scan recognition method provided in this embodiment can control the target device to perform plain scanning and arterial phase scanning and dynamic feature recognition in response to a scan instruction, to obtain first scan data and first dynamic features corresponding to the first scan data. If the first dynamic features do not match preset internal leakage features, the target device is controlled to perform delayed phase scanning and dynamic feature recognition, to obtain second scan data and second dynamic features corresponding to the second scan data. Based on the first dynamic features or the second dynamic features, a scan recognition result is determined. By performing multi-phase scanning and dynamic feature analysis, combining the matching judgment of the preset internal leakage features and the conditional delayed scanning strategy, and utilizing the complementarity of multi-time phase image data for feature fusion and classification modeling, the limitations of single arterial phase scanning can be compensated for, the dynamic feature comparison and analysis can distinguish different internal leakage types, and the image registration and feature fusion can reduce radiation exposure and artificial errors, thereby improving the scanning recognition accuracy of subtle features.

[0066] In one of the embodiments, the target device is a multi-slice spiral CT device, a dual-source CT device, or a magnetic resonance imaging device.

[0067] The multi-slice spiral CT device can be a computed tomography device that simultaneously collects data through multiple rows of detectors, thereby realizing high-speed rotation scanning and thin-layer imaging. The contrast agent visualization effect can be optimized by adjusting the tube voltage and pitch parameters. For example, the multi-slice spiral CT device can include a device that generates continuous tomographic images in a volume scanning mode.

[0068] The dual-source CT device can be a CT system equipped with two sets of X-ray tubes and detectors, thereby enabling simultaneous dual-energy imaging or shortening the scanning period. For example, the dual-source CT device can include a device that synchronously acquires arterial phase and delayed phase partial data.

[0069] The magnetic resonance imaging device can be a device that generates images of internal structures of the human body using a strong magnetic field and radio frequency pulses. Since there is no ionizing radiation and the soft tissue contrast is high, dynamic enhancement imaging can be used to monitor the flow and leakage of contrast agents in real time. For example, the magnetic resonance imaging device can be a device that combines T1-weighted imaging and parameter mapping technology.

[0070] In one specific embodiment, a dual-source CT device distinguishes intravascular and extravascular contrast agent signals through dual-energy imaging, combined with a material separation algorithm to extract iodine-based distribution characteristics, which can assist in the classification of internal leakage; a magnetic resonance imaging device can evaluate the degree of chronic extravasation or tissue infiltration by calculating parameters such as contrast agent arrival time, peak concentration, and washout rate.

[0071] The scanning and identifying method provided in this embodiment can achieve the technical effects of improving temporal resolution, spatial resolution, and lesion classification accuracy through the synergistic effect of device characteristics and multi-phase scanning, by selecting a multi-slice spiral CT device to achieve rapid full-vessel coverage scanning at the arterial phase and reduce motion artifacts, selecting a dual-source CT device to simultaneously acquire multi-phase data and distinguish contrast agent distribution characteristics, and selecting a magnetic resonance imaging device to provide soft tissue interaction information and analyze contrast agent diffusion patterns based on its non-radiation characteristics.

[0072] In one of the embodiments, the target device is controlled to perform plain scanning and arterial phase scanning and dynamic feature identification, to obtain first scanning data and first dynamic features corresponding to the first scanning data, which include:

[0073] The target device is controlled to perform plain scanning and arterial phase scanning, to obtain first scanning data; the first scanning data includes plain scanning data and arterial phase scanning data;

[0074] The plain scanning data and the arterial phase scanning data are registered based on a preset registration algorithm, to obtain first registered scanning data;

[0075] Dynamic features are extracted based on the first registered scanning data, to obtain first dynamic features.

[0076] The plain scanning data can be anatomical structure basic image data acquired by the medical imaging device under the condition of no contrast agent injection, and can be directly generated through the conventional scanning mode of the target device. For example, the plain scanning data can include original grayscale images of X-ray computed tomography (CT) or magnetic resonance imaging (MRI).

[0077] The arterial phase scanning data can be enhanced image data collected by the target device when the contrast agent reaches the target blood vessel region, and its generation depends on the specific time window scanning after the injection of the contrast agent. For example, the arterial phase scanning data can include high-contrast images of arterial vessel visualization.

[0078] The preset registration algorithm can be a mathematical model for eliminating spatial deviations of multi-phase scanning data, and the registration algorithm can include one or more of rigid registration, elastic registration, or feature-driven registration, but is not limited to these. For example, rigid registration can be achieved through translation and rotation transformation, elastic registration can be based on a thin plate spline or B-spline model, and feature-driven registration can use anatomical landmarks such as vessel bifurcation points or bone contours as alignment references.

[0079] The plain scan data is registered with the arterial phase scan data. Exemplarily, common features such as edges or texture information of the two sets of data can be extracted, a spatial transformation model is then established to calculate geometric offset parameters, and finally an iterative optimization algorithm is used to minimize an image similarity measure function to adjust the parameters, so as to eliminate the spatial deviation caused by patient body position movement or respiratory motion.

[0080] Dynamic feature extraction is performed based on the first registered scan data. Exemplarily, blood vessel segmentation combined with parameter analysis can be used. In one exemplary embodiment, threshold segmentation, region growing or deep learning model can be used to identify the blood vessel region in the registered arterial phase image, and the anatomical structure information in the plain scan data is used to locate the blood vessel wall or blood flow path, and then the parameters such as blood vessel wall integrity, blood flow signal intensity distribution and contrast agent flow velocity are extracted.

[0081] The scanning and recognition method provided by the embodiment can improve the spatio-temporal consistency of multi-phase scanning data, reduce feature extraction errors caused by target object activity or equipment errors, and improve the scanning and recognition accuracy of subtle features.

[0082] In one embodiment, the target device is controlled to perform delayed phase scanning and dynamic feature recognition, and the second scan data and the second dynamic feature corresponding to the second scan data are obtained, which further comprises:

[0083] The target device is controlled to perform delayed phase scanning to obtain second scan data; the second scan data comprises delayed phase scan data;

[0084] The delayed phase scan data is registered with the arterial phase scan data based on a preset registration algorithm to obtain second registered scan data;

[0085] Dynamic feature extraction is performed based on the second registered scan data to obtain second dynamic feature.

[0086] The delayed phase scan data can be anatomical structure image data collected by a medical imaging device within a specific time window, or can be obtained by controlling the target device to scan at a preset delay time point after contrast agent injection. Exemplarily, the delayed phase scan data can include CT, MRI and other devices capturing tomographic image sequences in the delayed phase.

[0087] The arterial phase scanning data is an anatomic image data obtained by arterial phase scanning in the first scanning data. The preset registration algorithm is an algorithm for realizing spatial alignment of multi-phase images, and is used to eliminate image spatial deviation caused by factors such as patient respiratory motion, device position fine adjustment or natural displacement of organs. An exemplary preset registration algorithm can include rigid registration (such as affine transformation), non-rigid registration (such as thin plate spline or B-spline model) or registration method based on feature point matching, and can also be other preset registration algorithms, which are not limited in the embodiment.

[0088] Based on the preset registration algorithm, the delay phase scanning data is registered with the arterial phase scanning data. Exemplarily, common anatomic landmark points or feature regions of the arterial phase and the delay phase images can be extracted as a reference framework, the spatial coordinates of the delay phase image are adjusted through iterative calculation to achieve the best matching of the anatomic structures of the arterial phase image at the pixel or voxel level, and finally a registration matrix containing spatial transformation parameters is generated to correct the positional deviation of the delay phase image. By eliminating image displacement or deformation, the problem of inaccurate feature positioning caused by registration error in traditional multi-phase scanning can be solved. For example, organ displacement caused by patient respiration can cause several millimeter deviation of the blood vessel position of the arterial phase and the delay phase images.

[0089] The dynamic feature extraction can be based on the second registration scanning data after registration, and the change of the imaging feature of the same anatomic position at different time points is quantified through image analysis algorithm. Exemplarily, the dynamic feature can include a density change slope, a region growing segmentation result or an edge detection parameter. For example, type I internal leakage may only show slight thickening of the local blood vessel wall in the arterial phase, while the density of the region may gradually increase due to continuous penetration of the contrast agent in the delay phase. Through superposition analysis of the registered images, the change slope of the density over time can be quantified, so as to distinguish low-pressure internal leakage from other types of lesions.

[0090] The scanning and identifying method provided in the embodiment can obtain delay phase scanning data by controlling the target device to capture the time sequence feature of the lesion, eliminate the spatial deviation of multi-phase images by using the preset registration algorithm and generate accurate second registration scanning data, and quantize the imaging change rule based on the registered data for dynamic feature extraction, so as to achieve the technical effects of significantly reducing image registration error and improving multi-phase feature contrast accuracy.

[0091] In one of the embodiments, the dynamic feature extraction based on the second registration scanning data obtains a second dynamic feature, which includes:

[0092] Based on the second registration scanning data, the CT value difference between the delay phase and the arterial phase is calculated to obtain an HU difference value;

[0093] Based on the scanning time difference of the second registration scanning data, the concentration change per unit time is calculated to obtain an accumulation rate;

[0094] calculating a gray scale distribution complexity based on the second registration scan data and a preset voxel window, to obtain a morphological entropy value;

[0095] generating a sensitivity matrix based on the Euclidean distance of the voxels in the second registration scan data to the target component, to obtain a spatial weight;

[0096] generating a shape feature map based on the HU difference value, the accumulation rate, the morphological entropy value and the spatial weight;

[0097] inputting the arterial phase scan data and the shape feature map into a pre-trained neural network to obtain a second dynamic feature.

[0098] The HU difference value can be the CT value difference of the same voxel position between the delay phase and the arterial phase, can be the CT value (HU delay ) of the delay phase minus the CT value (HU artery ) of the arterial phase, and is used to reflect the concentration change trend of the contrast agent in the lesion area. For example, this parameter can be obtained by voxel-by-voxel subtraction operation, and a difference feature map is generated to highlight the dynamic change area. In a specific embodiment, this parameter can be used to quantify the deposition difference of the developing agent at different time points, for example, the low-pressure leakage area can present a sustained HU value increase in the delay phase due to the slow exudation of the contrast agent.

[0099] The accumulation rate can be the ΔHU value change rate per unit time, which can be obtained by dividing the HU difference value ΔHU by the scanning time difference (T delay -T artery ) between the arterial phase and the delay phase, and is used to evaluate the penetration speed of the contrast agent in the lesion area, for example, type II internal leakage can cause a faster accumulation rate due to extravascular leakage, while type I internal leakage can present a stable rate due to continuous blood flow impact. For example, by combining the scanning time parameter, the accumulation rate can be calculated by linear regression or difference quotient method, and a rate distribution map is generated to distinguish different hemodynamic characteristics.

[0100] The morphological entropy value can be a quantitative indicator based on the local gray scale distribution complexity of the voxel, which can be the entropy of the probability of each gray value within a preset voxel window (such as 5x5x5), and is used to represent the texture heterogeneity of the tissue or lesion area. For example, the internal leakage area can present a higher entropy value due to the uneven accumulation of the contrast agent, while the normal tissue has a lower entropy value due to uniform absorption. For example, the calculation of this parameter can be to calculate the frequency distribution p i of each gray value within the statistical window, and substitute it into the entropy formula (entropy = -Σ(p i x log2(p i )) to calculate.

[0101] The spatial weight can be a sensitivity matrix generated by the Euclidean distance from the voxel to the target component (such as a stent), and the weight = 1 / (1+d), wherein d is the distance value, and further, the weight value can be doubled when d≤50mm. The spatial weight is used to highlight the voxel features close to the stent area, and the internal leakage often occurs at the stent adhesion, and the sensitivity is higher as the distance is closer. Exemplarily, the three-dimensional coordinates of the stent can be extracted, the shortest distance from each voxel to the surface of the stent is calculated, and the weight coefficient is calculated according to the distance value by substituting the formula to generate the spatial sensitivity weight map.

[0102] The shape feature map can be a comprehensive feature representation after fusing the above four feature parameters, and exemplarily, can include one or more of weighted summation, feature map superposition, channel concatenation and the like. The generation of the shape feature map can integrate multi-dimensional dynamic information to form a stereoscopic feature description of the internal leakage area, for example, the voxel with high ΔHU, high accumulation rate, high entropy value and close to the stent is more likely to belong to the internal leakage area.

[0103] Exemplarily, the pre-trained neural network can be a 3D-Unet architecture, containing 4 levels of down-sampling and up-sampling modules, and fusing multi-scale features through a jump connection. The use of the network is to perform pixel-level segmentation on complex three-dimensional structures, such as distinguishing internal leakage areas from surrounding tissues. Exemplarily, the input mode of the neural network can adopt a dual-channel design: the original image of the arterial phase provides an anatomical structure reference, and the ΔHU feature map provides dynamic change information, which together constitute the network input. The input voxel block size is 128×128×128, which can ensure sufficient context information while taking into account the computational efficiency.

[0104] Further, after inputting into the neural network, it can include multi-condition decision output, including threshold segmentation and morphological post-processing, such as setting the probability threshold to 0.5 to binarize the result, and removing small area noise or connecting broken areas through erosion, dilation and the like to improve the smoothness and integrity of the segmentation result.

[0105] The scanning and identifying method provided in the embodiment can quantify the contrast agent deposition change by calculating the CT value difference between the delay phase and the arterial phase, analyze the contrast agent penetration speed by combining the scanning time difference, evaluate the tissue heterogeneity by using the local gray scale distribution complexity, generate the spatial sensitivity weight based on the Euclidean distance, fuse the multi-dimensional features to generate the shape feature map, and input into the 3D-Unet network to realize the pixel-level segmentation and morphological post-processing, which can achieve the technical effects of accurately capturing the contrast agent time sequence change rule, enhancing the identification ability of small lesions, focusing on the clinically concerned area, and improving the segmentation accuracy and robustness.

[0106] In one of the embodiments, the target device is controlled to perform the delay phase scanning and the dynamic feature recognition, to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data, which comprises:

[0107] Based on the scanning instruction and the object information of the target object, the scanning parameter of the target device is determined;

[0108] Based on the scanning parameter, the target device is controlled to perform the delay phase scanning, to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data.

[0109] The object information of the target object can be a set of objective data related to the scanned object, such as patient size (e.g. height, weight), physiological state (e.g. heart rate, breathing rate), medical history, preliminary anatomical structure features of the current scanning area (e.g. vessel diameter obtained in the arterial phase scanning, lesion location, etc.), which can be input by an input device or automatically analyzed from the previous scanning data. For example, the object information can include patient BMI value, arterial phase vessel visualization degree, etc.

[0110] The scanning parameter can be a physical or technical parameter of the CT device that can be adjusted during scanning, including but not limited to tube voltage, tube current, scanning slice thickness, pitch, reconstruction algorithm type, scanning time window, etc., which can affect the spatial resolution, density resolution, radiation dose and scanning efficiency of the scanning image. In this embodiment, the scanning parameter can be a set of dynamically adapted CT device operation parameters, such as tube current value adjusted according to the size of the target object, scanning slice thickness optimized according to the size of the lesion area.

[0111] By receiving the parameter information contained in the scanning instruction and the key features extracted from the object information, a comprehensive analysis can be performed in combination with a preset algorithm or an artificial intelligence model to determine the scanning parameter. For example, the optimal tube voltage value is predicted by a regression model to balance image noise and radiation dose, or the starting time of the delay phase scanning is dynamically calculated according to the arterial phase contrast agent peak time. For example, if the arterial phase shows that the contrast agent stays in a certain area for a long time, the delay phase scanning interval can be extended to capture the longer exudation process, and a customized scanning parameter configuration can be generated, which can include the starting time of the delay phase scanning, the scanning slice thickness, the selection of the iterative reconstruction algorithm, etc. Through adaptive adjustment of the parameter, the delay phase scanning can be more accurately focused on the dynamic change features of the target area.

[0112] The second scan data can be image or signal data collected in a delay phase scanning process, and can be generated after the target device performs customized parameter configuration. Correspondingly, the second dynamic feature can be a quantitative index reflecting the dynamic change of the target object in the delay phase scanning period, such as a contrast agent excretion path, a lesion area density change trend, etc.

[0113] The scanning and identifying method provided in this embodiment can dynamically generate personalized scanning parameters based on a scanning instruction and object information, combine delay phase scanning driven by parameter configuration and dynamic feature extraction, and through parameter adaptation, can make the delay phase scanning accurately match the physiological features and lesion characteristics of the target object, for example, dynamically adjust a delay time window according to the arterial phase result to capture details of dynamic distribution of the contrast agent, etc., through reducing the need for manual intervention, make the multi-phase scanning process more automated, shorten the identification time and reduce human operation errors, so as to achieve the technical effect of improving the identification accuracy.

[0114] In one of the embodiments, the second scan data includes first delay phase scan data and second delay phase scan data; and the control of the target device to perform delay phase scanning and dynamic feature identification to obtain the second scan data and the second dynamic feature corresponding to the second scan data further includes:

[0115] The control of the target device to perform delay phase scanning and dynamic feature identification to obtain the first delay phase scan data and the first delay phase dynamic feature corresponding to the first delay phase scan data;

[0116] If the first delay phase dynamic feature does not match the preset internal leakage feature, the control of the target device to perform delay phase scanning and dynamic feature identification to obtain the second delay phase scan data and the second delay phase dynamic feature corresponding to the second delay phase scan data;

[0117] Based on the first delay phase dynamic feature and the second delay phase dynamic feature, the second dynamic feature is determined.

[0118] The first delay phase scan data can be CT image data obtained by first delay phase scanning after arterial phase scanning, can be obtained by scanning 5-10 minutes after the peak of the arterial phase, and the delay time is set based on the kinetic characteristics of the excretion of the contrast agent to the extravascular space, which can preliminarily show the early imaging features of small lesions such as low-pressure internal leakage.

[0119] The first delay phase dynamic feature can be a set of parameters such as delay early contrast agent distribution mode and local concentration change rate, which contains quantitative indexes such as regional concentration gradient and time-density curve slope, and can be extracted and characterized by an image processing algorithm to represent the initial dynamic behavior of the lesion.

[0120] The preset internal leakage feature can be a threshold or a morphological criterion for determining whether the first scan result meets the identification condition. Examples include a specific concentration threshold, an abnormal signal diffusion rate, or a degree of morphological abnormality, which are used to determine whether the subsequent scanning process needs to be started.

[0121] The second delay period scan data can be CT image data obtained by further extending the delay time (e.g., 10-15 minutes) after the first delay period scan. The delay time is designed based on the difference in penetration kinetics of contrast agents in different internal leakage types. Examples include a delay imaging window that covers complex lesions such as type III internal leakage.

[0122] The second delay period dynamic feature can be the distribution pattern and concentration change parameters of the contrast agent extracted in the second delay period scan. Examples include the spatial expansion rate of the exudation area and the concentration decay curve, which are used to supplement the dynamic evolution information that is not captured in the first scan.

[0123] Controlling the target device to perform delay period scanning and dynamic feature recognition, examples can achieve adaptive adjustment of scanning strategy through dynamic threshold judgment mechanism, including: first performing the first delay period scan and extracting dynamic features, then comparing the feature parameters with the preset internal leakage features (such as whether the concentration is lower than the threshold or the morphology meets the abnormal standard), if not matching, triggering the second scan. This process avoids the radiation dose redundancy caused by fixed two scans through conditional branching process, for example, some cases can terminate the process after the first scan.

[0124] Determining the second dynamic feature based on the first delay period dynamic feature and the second delay period dynamic feature, examples can be achieved through time series modeling and difference analysis, including: feature fusion integrates the concentration change curve, exudation area expansion rate and other parameters of the two scans through weighted algorithm or deep learning model to form a multi-dimensional feature vector; difference analysis identifies the time-dependent features specific to internal leakage types by comparing the imaging differences of the two scans (such as the difference in concentration change slope of specific areas or the difference in abnormal signal diffusion range). For example, type II internal leakage may gradually decrease in concentration after the first scan, while type III internal leakage shows a continuous exudation feature. Such analysis can improve the discrimination ability of internal leakage classification.

[0125] The scanning and identifying method provided in the embodiment acquires preliminary dynamic features by performing first delay period scanning in stages, initiates second scanning to cover delay development lesions by combining dynamic threshold judgment conditions, constructs a multi-dimensional dynamic feature model by time sequence feature fusion and difference analysis, reduces average radiation dose by adaptive scanning strategy, and terminates the process after the first scanning for some cases, so that the effectiveness of matching internal leakage features can be improved, the false detection rate of dynamic features can be further reduced based on the relevance analysis of dynamic features at multiple time points, and the technical effects of maintaining the identification efficiency and accurately capturing the dynamic evolution law of subtle lesions are achieved.

[0126] In order to more clearly set forth the technical solutions of the present application, a detailed embodiment is further provided.

[0127] In one embodiment, as shown in Figure 3 A scanning and identifying method is provided, comprising:

[0128] Device and scanning preparation: a multi-layer spiral CT device with multi-phase scanning function is selected to ensure stable device performance and image resolution meeting detection requirements, for example, a CT machine with 64 rows or more detectors, whose spatial resolution can reach sub-millimeter level. Before scanning, the patient needs to lie on the examination bed and keep the body still to avoid motion artifacts interfering with image quality. Through breathing training, the patient is guided to perform regular breathing or breath holding during scanning to ensure the consistency of the same layer position in different phase scanning.

[0129] First, conventional plain scanning and conventional arterial phase scanning are performed, and image registration and neural network are used for internal leakage detection. If dynamic features corresponding to the preset internal leakage features are found, the scanning is terminated; otherwise, first delay period scanning is performed: the target object position is kept unchanged, and the other parameters are the same as those in the arterial phase. Under the condition of ensuring that the delay period contrast agent concentration is maintained at an appropriate level, the delay period image is acquired, and the scanning time can be 30 to 45 seconds.

[0130] In one specific embodiment, the age, BMI, and suspected internal leakage position of each target object can be automatically acquired according to the Dicom label, and the scanning time is calculated.

[0131] If the first delay scanning is performed, image registration and neural network are used for internal leakage detection. If dynamic features corresponding to the preset internal leakage features are found, the scanning is terminated; otherwise, second delay period scanning is performed, the target object position is kept unchanged, and the other parameters are the same as those in the first delay scanning. Whether or not the suspected internal leakage is detected, the scanning is terminated.

[0132] In the embodiment, the image registration and neural network detection can include:

[0133] (1) Image registration preprocessing

[0134] The arterial phase and delayed phase DICOM images were aligned in three-dimensional space using the Demons non-rigid registration algorithm (100 iterations, standard deviation parameter 1.5). By calculating the displacement field, the anatomical structure displacement caused by respiratory motion was eliminated (average registration error ≤0.8 mm), ensuring that the voxels at the same anatomical position corresponded accurately.

[0135] (2) Dynamic feature extraction

[0136] The voxel-level feature calculation was performed on the registered image pair, and the features included the following contents:

[0137] ΔHU value: Calculate the CT value difference between the delayed phase and the arterial phase (ΔHU = HU delay -HU artery ).

[0138] Accumulation rate: Calculate the concentration change per unit time according to the scanning time difference, the calculation formula can be: rate = ΔHU / (T delay -T artery ).

[0139] Morphological entropy value: Calculate the complexity of gray scale distribution in a 5x5x5 voxel window, the calculation formula can be: entropy = -Σ(p i x log2(p i )).

[0140] Spatial weight: Based on the Euclidean distance from the voxel to the stent, a sensitivity matrix is generated, the calculation formula can be: weight = 1 / (1+d), where d≤50mm, the weight is multiplied by 2.

[0141] (3) 3D-Unet neural network recognition

[0142] The arterial phase image and ΔHU feature map are input as a dual-channel 128x128x128 voxel block, and processed through a 4-level encoding-decoding structure.

[0143] (4) Multi-condition decision output

[0144] Threshold segmentation and morphological filtering are performed on the probability map to obtain the detection result of the region of interest, which is used as a reference to assist the user in making corresponding judgments.

[0145] In some other embodiments, the scanning device can also use dual-source CT for scanning. Dual-source CT has two sets of X-ray sources and detector systems, which can complete different energy or different phase scanning in a very short time, and can reduce artifacts caused by patient's breathing, heartbeat and other slight movements. Material separation is performed using high and low energy data, which better identifies contrast agents, and tube current is automatically adjusted according to image noise. The contrast agent injection scheme and image analysis method are similar to the above-mentioned multi-slice spiral CT process, and can also obtain arterial phase and delayed phase images for endoleak detection.

[0146] In some other embodiments, for patients with renal dysfunction or allergy to iodine contrast agents, MRI enhanced scanning can also be used, combined with special contrast agents (such as gadolinium contrast agents) to replace CT scanning. Using the high resolution and multi-parameter imaging advantages of MRI, the signal change characteristics of the endoleak area are observed for detection.

[0147] As shown in Figure 4 and Figure 6 , respectively, the arterial phase scanning data of the present embodiment is shown, and it can be seen that thrombosis occurs around the stent, but there is no obvious endoleak at the proximal end of the stent, as shown in Figure 5 and Figure 7 , respectively, the delayed phase scanning data of the present embodiment is shown, and according to the arrows shown, it can be seen that the false lumen is connected with the left subclavian artery, and the dynamic characteristics of the type II endoleak can be detected by the endoleak detection algorithm based on neural networks. Figure 8 and Figure 9 are images after three-dimensional reconstruction, and through the reconstructed images, the source reference of the type II endoleak can be more clearly shown, which can assist the user to make further judgments.

[0148] The scanning and identifying method provided by the embodiment can capture the dynamic distribution change of the contrast agent at different time phases after adding the delayed phase scanning, and the low-tension internal leakage can be more clearly displayed due to further penetration and accumulation of the contrast agent in the delayed phase, so that the internal leakage that is difficult to be found in the arterial phase can be displayed, thereby improving the detection rate of the internal leakage. By comparing the images at different time phases, the display characteristic difference of the internal leakage at different time nodes can be observed. For example, the type I internal leakage can present a certain specific shape and blood flow signal in the arterial phase due to the special hemodynamic characteristics of the adhesion part of the stent, and in the delayed phase, the type I internal leakage can present a change rule different from other types such as the type II internal leakage due to the continuous blood flow impact and different dispersion modes of the contrast agent. The user can accurately distinguish the internal leakage types according to these multi-phase image characteristics, and trace the source blood vessels or interstitial spaces of the internal leakage according to the position where the contrast agent first appears abnormal concentration or flow anomaly, thereby providing a key basis for subsequent work. Compared with the previous vascular ultrasound or single arterial phase MSCTA examination, the accuracy of the internal leakage typing of the method is higher, and the accuracy of the source judgment is also higher, which can greatly improve the detection accuracy. Compared with DSA, the CT multi-phase scanning can avoid additional physical trauma, infection risk and psychological pressure in the follow-up process, and the operation is more simple, which can shorten the examination time, reduce the medical cost, and make the patient more easily accept the regular postoperative review, so as to help to find potential internal leakage problems in time and ensure the long-term reliability of the aortic stent.

[0149] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0150] Based on the same inventive concept, the embodiment of the present application also provides a scanning and identifying device for implementing the scanning and identifying method as described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, and therefore the specific limitations in one or more scanning and identifying device embodiments provided below can refer to the limitations of the scanning and identifying method described above, which will not be described here again.

[0151] In one embodiment, as shown in Figure 10 A scanning and identifying apparatus is provided, which comprises:

[0152] A first scanning module 100 is configured to control a target device to perform plain scan and arterial phase scan and dynamic feature identification respectively in response to a scanning instruction, to obtain first scanning data and a first dynamic feature corresponding to the first scanning data.

[0153] A second scanning module 200 is configured to control the target device to perform delay phase scan and dynamic feature identification if the first dynamic feature does not match a preset internal leakage feature, to obtain second scanning data and a second dynamic feature corresponding to the second scanning data.

[0154] A result determining module 300 is configured to determine a scanning and identifying result based on the first dynamic feature or the second dynamic feature.

[0155] In some embodiments, the target device is a multi-slice spiral CT device, a dual-source CT device or a magnetic resonance imaging device.

[0156] In some embodiments, the first scanning module 100 is further configured to:

[0157] control the target device to perform plain scan and arterial phase scan respectively to obtain first scanning data; the first scanning data comprises plain scan data and arterial phase scan data;

[0158] register the plain scan data and the arterial phase scan data based on a preset registration algorithm to obtain first registered scanning data;

[0159] extract a dynamic feature based on the first registered scanning data to obtain the first dynamic feature.

[0160] In some embodiments, the second scanning module 200 is further configured to:

[0161] control the target device to perform delay phase scan to obtain second scanning data; the second scanning data comprises delay phase scan data;

[0162] register the delay phase scan data and the arterial phase scan data based on a preset registration algorithm to obtain second registered scanning data;

[0163] extract a dynamic feature based on the second registered scanning data to obtain the second dynamic feature.

[0164] In some embodiments, the second scanning module 200 is further configured to:

[0165] calculate a CT value difference between the delay phase and the arterial phase based on the second registered scanning data to obtain a HU difference.

[0166] calculating a concentration change per unit time based on a scan time difference of the second registration scan data, to obtain an accumulation rate;

[0167] calculating a gray scale distribution complexity based on the second registration scan data and a preset voxel window, to obtain a morphological entropy value;

[0168] generating a sensitivity matrix based on a Euclidean distance from a voxel to a target component in the second registration scan data, to obtain a spatial weight;

[0169] generating a shape feature map based on the HU difference, the accumulation rate, the morphological entropy value and the spatial weight;

[0170] inputting the arterial phase scan data and the shape feature map into a pre-trained neural network, to obtain the second dynamic feature.

[0171] In some embodiments, the second scanning module 200 is further configured to:

[0172] determining a scanning parameter of the target device based on the scanning instruction and object information of the target object;

[0173] controlling the target device to perform a delay phase scan based on the scanning parameter, to obtain second scan data and a second dynamic feature corresponding to the second scan data.

[0174] In some embodiments, the second scan data includes first delay phase scan data and second delay phase scan data; the second scanning module 200 is further configured to:

[0175] controlling the target device to perform a delay phase scan and dynamic feature recognition, to obtain first delay phase scan data and a first delay phase dynamic feature corresponding to the first delay phase scan data;

[0176] if the first delay phase dynamic feature does not match a preset internal leakage feature, controlling the target device to perform a delay phase scan and dynamic feature recognition, to obtain second delay phase scan data and a second delay phase dynamic feature corresponding to the second delay phase scan data;

[0177] determining the second dynamic feature based on the first delay phase dynamic feature and the second delay phase dynamic feature.

[0178] Each module in the above scanning and recognition apparatus can be realized by software, hardware and combinations thereof, in whole or in part. Each module can be embedded in or independent of a processor in a computer device in hardware form, or stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform operations corresponding to each module.

[0179] In one embodiment, a computer device is provided, which can be a terminal, and an internal structure diagram thereof can be as shown in FIG. 1. Figure 11 The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement a scanning and identification method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0180] Those skilled in the art can understand that Figure 11 The structure shown in FIG. 1 is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0181] In one embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the scanning and identification method of any of the above embodiments.

[0182] In response to the scanning instruction, the target device is controlled to perform plain scan and arterial phase scan and dynamic feature identification, to obtain first scanning data and first dynamic features corresponding to the first scanning data;

[0183] If the first dynamic features do not match the preset internal leakage features, the target device is controlled to perform delay phase scan and dynamic feature identification, to obtain second scanning data and second dynamic features corresponding to the second scanning data;

[0184] Based on the first dynamic features or the second dynamic features, a scanning and identification result is determined.

[0185] In one embodiment, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the scanning and identification method of any of the above embodiments.

[0186] In response to the scanning instruction, the target device is controlled to perform plain scan and arterial phase scan and dynamic feature recognition respectively, to obtain first scanning data and first dynamic feature corresponding to the first scanning data;

[0187] If the first dynamic feature does not match the preset internal leakage feature, the target device is controlled to perform delay phase scan and dynamic feature recognition, to obtain second scanning data and second dynamic feature corresponding to the second scanning data;

[0188] Based on the first dynamic feature or the second dynamic feature, a scanning recognition result is determined.

[0189] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0190] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0191] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0192] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method of scan recognition, the method comprising: The scanning recognition method comprises: In response to a scanning instruction, the target device is controlled to perform plain scan and arterial phase scan and dynamic feature recognition respectively, to obtain first scanning data and first dynamic features corresponding to the first scanning data; the control of the target device to perform plain scan and arterial phase scan and dynamic feature recognition respectively to obtain first scanning data and first dynamic features corresponding to the first scanning data comprises: the target device is controlled to perform plain scan and arterial phase scan respectively to obtain first scanning data; the first scanning data comprises plain scan data and arterial phase scan data; the plain scan data and the arterial phase scan data are registered based on a preset registration algorithm to obtain first registered scanning data; and dynamic features are extracted based on the first registered scanning data to obtain the first dynamic features; If the first dynamic features do not match preset internal leakage features, the target device is controlled to perform delay phase scan and dynamic feature recognition to obtain second scanning data and second dynamic features corresponding to the second scanning data; the control of the target device to perform delay phase scan and dynamic feature recognition to obtain second scanning data and second dynamic features corresponding to the second scanning data further comprises: the target device is controlled to perform delay phase scan to obtain second scanning data; the second scanning data comprises delay phase scan data; the delay phase scan data and the arterial phase scan data are registered based on a preset registration algorithm to obtain second registered scanning data; and dynamic features are extracted based on the second registered scanning data to obtain the second dynamic features; the extraction of the second dynamic features based on the second registered scanning data comprises: a CT value difference between the delay phase and the arterial phase is calculated based on the second registered scanning data to obtain a HU difference value; a concentration change per unit time is calculated based on a scanning time difference of the second registered scanning data to obtain an accumulation rate; a gray scale distribution complexity is calculated based on the second registered scanning data and a preset voxel window to obtain a morphological entropy value; a sensitivity matrix is generated based on a Euclidean distance of a voxel to a target component in the second registered scanning data to obtain a spatial weight; a shape feature map is generated based on the HU difference value, the accumulation rate, the morphological entropy value and the spatial weight; and the arterial phase scan data and the shape feature map are input into a pre-trained neural network to obtain the second dynamic features; The scanning recognition result is determined based on the first dynamic features or the second dynamic features.

2. The method of claim 1, wherein, The target device is a multi-slice CT device, a dual-source CT device or a magnetic resonance imaging device.

3. The method of claim 1, wherein, The control of the target device to perform delay phase scan and dynamic feature recognition to obtain second scanning data and second dynamic features corresponding to the second scanning data comprises: Based on the scanning instruction and object information of a target object, scanning parameters of the target device are determined; The target device is controlled to perform delay phase scan based on the scanning parameters to obtain second scanning data and second dynamic features corresponding to the second scanning data.

4. The method of claim 1, wherein, The second scan data includes first delay period scan data and second delay period scan data; and the control of the target device to perform delay period scanning and dynamic feature recognition to obtain second scan data and second dynamic features corresponding to the second scan data further includes: controlling the target device to perform delay period scanning and dynamic feature recognition to obtain first delay period scan data and first delay period dynamic features corresponding to the first delay period scan data; if the first delay period dynamic features do not match the preset internal leakage features, controlling the target device to perform delay period scanning and dynamic feature recognition to obtain second delay period scan data and second delay period dynamic features corresponding to the second delay period scan data; determining the second dynamic features based on the first delay period dynamic features and the second delay period dynamic features.

5. A scanning recognition device, characterized in that The scan recognition device includes: a first scan module configured to, in response to a scan instruction, control a target device to perform plain scanning and arterial phase scanning and dynamic feature recognition respectively to obtain first scan data and first dynamic features corresponding to the first scan data; the control of the target device to perform plain scanning and arterial phase scanning and dynamic feature recognition respectively to obtain first scan data and first dynamic features corresponding to the first scan data includes: controlling the target device to perform plain scanning and arterial phase scanning respectively to obtain first scan data; the first scan data includes plain scan data and arterial phase scan data; performing registration on the plain scan data and the arterial phase scan data based on a preset registration algorithm to obtain first registered scan data; and performing dynamic feature extraction based on the first registered scan data to obtain the first dynamic features; The second scanning module is configured to, if the first dynamic feature does not match the preset internal leakage feature, control the target device to perform a delay period scanning and dynamic feature recognition, to obtain second scanning data and a second dynamic feature corresponding to the second scanning data. The control of the target device to perform the delay period scanning and the dynamic feature recognition to obtain the second scanning data and the second dynamic feature corresponding to the second scanning data further includes: control of the target device to perform the delay period scanning to obtain the second scanning data. The second scanning data includes delay period scanning data. The delay period scanning data is registered with the arterial period scanning data based on a preset registration algorithm to obtain second registered scanning data. Dynamic feature extraction is performed based on the second registered scanning data to obtain the second dynamic feature. The dynamic feature extraction based on the second registered scanning data to obtain the second dynamic feature includes: calculation of a CT value difference between the delay period and the arterial period based on the second registered scanning data to obtain a HU difference value. Calculation of a concentration change per unit time based on a scanning time difference of the second registered scanning data to obtain an accumulation rate. Calculation of a gray scale distribution complexity based on the second registered scanning data and a preset voxel window to obtain a morphological entropy value. Generation of a sensitivity matrix based on a Euclidean distance of a voxel to a target component in the second registered scanning data to obtain a spatial weight. Generation of a shape feature map based on the HU difference value, the accumulation rate, the morphological entropy value, and the spatial weight. Input of the arterial period scanning data and the shape feature map into a pre-trained neural network to obtain the second dynamic feature. The result determination module is configured to determine a scanning recognition result based on the first dynamic feature or the second dynamic feature. 6.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-5 when the computer program is executed by the processor. The processor executes the computer program to implement the method in any one of claims 1 to 4.

7. A computer readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to implement the method in any one of claims 1 to 4. The computer program is executed by the processor to implement the method in any one of claims 1 to 4.

Images