Angiographic image generation method, apparatus, medical imaging device, and storage medium

Abstract

The application discloses a method and device for generating a subtraction image, a medical imaging device and a storage medium. First, a target contrast image and a preliminary matching mask image matching the motion law of the target contrast image are obtained. Then, the preliminary matching mask image is processed to obtain a target matching mask image with a background similar to that of the target contrast image. Finally, a subtraction image corresponding to the target contrast image is generated based on the target matching mask image and the target contrast image. By obtaining a preliminary matching mask image matching the motion law of the target contrast image, the method can better cope with low frame rate acquisition or patient restlessness and the like. Generating a subtraction image corresponding to the target contrast image based on the target matching mask image with a background similar to that of the target contrast image and the target contrast image can help better eliminate background interference in the target contrast image, thereby improving the quality of the subtraction image.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method, apparatus, medical imaging device, and storage medium for generating vascular subtraction images. Background Technology

[0002] Digital subtraction angiography (DSA) uses digital processing and subtraction technology to highlight the contrast agent within blood vessels and improve the visibility of vascular structures by comparing X-ray images before and after contrast agent injection, thereby improving the accuracy of medical diagnosis.

[0003] During actual imaging, due to the patient's breathing, heartbeat, and other factors, the background of the X-ray images before and after contrast agent injection cannot be exactly the same, resulting in artifacts in the subtraction image that interfere with the display of blood vessels. Summary of the Invention

[0004] The embodiments described in this specification aim to at least partially solve one of the technical problems in the related art. To this end, the embodiments of this specification propose a method, apparatus, medical imaging device, and storage medium for generating vascular subtraction images.

[0005] This specification provides a method for generating vascular subtraction images, the method comprising:

[0006] Acquire the target contrast image and a preliminary matching mask image that matches the motion pattern of the target contrast image;

[0007] The preliminary matching mask image is processed to obtain a target matching mask image whose background is similar to that of the target imaging image;

[0008] Based on the target matching mask image and the target angiography image, a vascular subtraction image corresponding to the target angiography image is generated.

[0009] In one implementation, the preliminary matching mask image is determined by:

[0010] Obtain the mask image sequence;

[0011] The target phase stage of the target contrast image is determined based on the phase of the target contrast image and the phase of a specified set of contrast images; wherein at least a portion of the specified set of contrast images is located before the target contrast image, or the specified set of contrast images is located after the target contrast image;

[0012] The preliminary matching mask image is determined in the mask image sequence based on the target phase stage and the motion law data of the target imaging image.

[0013] In one implementation, the mask image sequence corresponds to multiple motion cycles; determining the preliminary matching mask image in the mask image sequence based on the target phase stage and the motion pattern data of the target angiographic image includes:

[0014] In the masked image sequence, determine multiple frames of masked images located within the target phase phase of each motion cycle;

[0015] In each motion cycle, among the multi-frame masked images located within the target phase phase, a set of reference masked images whose similar data to the specified angiographic image set satisfies preset similarity screening conditions is determined;

[0016] In the masked image sequence, a masked image that is located within the motion cycle of the reference masked image set and matches the motion pattern data of the target imaging image is identified as the preliminary matching masked image.

[0017] In one implementation, determining a set of reference mask images whose similarity data to the specified angiographic image set satisfies preset similarity screening conditions among the multi-frame mask images located within the target phase phase in each motion cycle includes:

[0018] Based on the motion pattern data of the specified angiographic image set, match the multi-frame mask images located in the target phase phase in each motion cycle to obtain the mask set to be screened in each motion cycle;

[0019] In each of the multiple motion cycles, the set of reference mask images that is most similar to the specified set of contrast images is determined.

[0020] In one implementation, the preliminary matching mask images are multiple images, and the multiple preliminary matching mask images constitute a preliminary matching mask set. The step of determining the preliminary matching mask images from the mask image sequence based on the target phase stage and the motion law data of the target imaging image includes:

[0021] Determine multiple frames of masked images located within the target phase phase in the masked image sequence;

[0022] Based on the motion pattern data of the target imaging image, matching is performed in multiple mask images located within the target phase phase to obtain the preliminary matching mask set.

[0023] In one implementation, the preliminary matching mask images are multiple images, which constitute a preliminary matching mask set. The target matching mask image is obtained by fusing the preliminary matching mask images in the preliminary matching mask set. The step of processing the preliminary matching mask images to obtain a target matching mask image with a background that is similar to the target contrast image includes:

[0024] The target matching mask image is obtained by locating key points and calculating pixel values ​​based on the preliminary matching mask images in the preliminary matching mask set.

[0025] In one implementation, the preliminary matching mask set includes a first mask image and a second mask image; the step of performing key point localization and pixel value calculation based on the preliminary matching mask images in the preliminary matching mask set to obtain the target matching mask image includes:

[0026] Determine the first position data corresponding to key points in the first masking image and the second position data corresponding to key points in the second masking image; wherein, the key points in the first masking image match the key points in the second masking image;

[0027] Interpolation processing is performed using the first phase data of the first mask image, the second phase data of the second mask image, the third phase data of the target imaging image, the first position data, and the second position data to obtain the target position data of key points in the target matching mask image;

[0028] Interpolation processing is performed on the pixel data of the first mask image, the pixel data of the second mask image, the first position data, and the second position data to obtain the pixel values ​​of key points in the target matching mask image;

[0029] The target matching mask image is obtained based on the target location data and pixel values ​​of key points in the target matching mask image.

[0030] In one implementation, the preliminary matching mask set has two frames of images; before determining the first location data corresponding to key points in the first mask image, the method further includes:

[0031] In the preliminary matching mask set, the preliminary matching mask image that is closest to the motion law data of the target imaging image is determined and used as the first mask image;

[0032] The preliminary matching mask images other than the first mask image in the preliminary matching mask set are used as the second mask image; wherein, the second position data is determined on the second mask image based on the first position data by template matching.

[0033] In one implementation, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image includes:

[0034] Registration is performed based on the target matching mask image and the target angiography image to obtain a registered mask image; the registered mask image is obtained by subtracting the registered mask image from the target angiography image; or

[0035] The target angiography image is obtained by subtracting the target matching mask image from the target angiography image.

[0036] In one implementation, the preliminary matching mask images are multiple images, and the multiple preliminary matching mask images constitute a preliminary matching mask set; the preliminary matching mask set is determined by the following method:

[0037] Obtain the mask image sequence;

[0038] Determine the third position data of the key points of the target imaging image and the fourth position data of the key points of each frame of the masked image in the masked image sequence;

[0039] Based on the third position data and the fourth position data, determine the image similarity data between each frame of the mask image in the mask image sequence and the target imaging image;

[0040] The preliminary matching mask set is determined in the mask image sequence based on the image similarity data.

[0041] This specification provides a vascular subtraction image generation apparatus, the apparatus comprising:

[0042] The image acquisition module is used to acquire the target contrast image and the preliminary matching mask image that matches the motion pattern of the target contrast image;

[0043] The image processing module is used to process the preliminary matching mask image to obtain a target matching mask image whose background is similar to that of the target imaging image;

[0044] The subtraction image generation module is used to generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image.

[0045] This specification provides a medical imaging device, which includes: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, which, when executed by the one or more processors, cause the one or more processors to perform the steps of the method described in any of the above embodiments.

[0046] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the above embodiments.

[0047] This specification provides a computer program product that includes instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0048] In the above-described embodiment, firstly, a target angiography image and a preliminary matching mask image that matches the motion pattern of the target angiography image are acquired. Then, the preliminary matching mask image is processed to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image. Finally, a vascular subtraction image corresponding to the target angiography image is generated based on the target matching mask image and the target angiography image. By acquiring a preliminary matching mask image that matches the motion pattern of the target angiography image, the requirements for mask image quality are reduced, better handling situations such as low frame rate acquisition or patient agitation, thereby improving the robustness and reliability of the vascular subtraction image generation method. Processing the preliminary matching mask image that matches the motion pattern of the target angiography image to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image increases the background similarity between the mask image used to generate the vascular subtraction image and the target angiography image, making the target matching mask image and the target angiography image visually more unified and realistic. Therefore, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image helps to better eliminate background interference in the target angiography image, thereby highlighting the shape and contour of the blood vessels and improving the quality of the vascular subtraction image. Attached Figure Description

[0049] Figure 1a A flowchart of the method for generating vascular subtraction images provided in the embodiments of this specification;

[0050] Figure 1b A schematic flowchart illustrating the vascular subtraction image generation method provided in the embodiments of this specification;

[0051] Figure 1c A schematic diagram of the respiratory gating output signal and its corresponding time provided for the embodiments of this specification;

[0052] Figure 2 A schematic flowchart illustrating the vascular subtraction image generation method provided in the embodiments of this specification;

[0053] Figure 3a A flowchart illustrating the process of determining a preliminary set of matching masks provided for the implementation of this specification;

[0054] Figure 3b A schematic diagram illustrating the determination of a preliminary matching mask set for embodiments of this specification;

[0055] Figure 4a A flowchart illustrating the process of determining a preliminary set of matching masks provided for the implementation of this specification;

[0056] Figure 4b A schematic diagram illustrating the determination of a preliminary matching mask set for embodiments of this specification;

[0057] Figure 5a A flowchart illustrating the process of determining a set of reference mask images for embodiments of this specification;

[0058] Figure 5b A schematic diagram illustrating the determination of a set of reference mask images for embodiments of this specification;

[0059] Figure 6 A flowchart illustrating the process of determining a preliminary set of matching masks provided for the implementation of this specification;

[0060] Figure 7 A schematic diagram illustrating the process of obtaining a target matching mask image provided for embodiments of this specification;

[0061] Figure 8a A flowchart illustrating the determination of the first mask image and the second mask image provided for embodiments of this specification;

[0062] Figure 8b A schematic diagram illustrating the determination of the first mask image and the second mask image for the purposes of embodiments of this specification;

[0063] Figure 9a A schematic diagram of a vascular subtraction image generated by registering a fixed mask image with a first frame angiography image, as provided in the embodiments of this specification.

[0064] Figure 9b A schematic diagram of a vascular subtraction image generated by registering a fixed mask image with a second angiographic image, as provided in the embodiments of this specification.

[0065] Figure 9cA schematic diagram of a vascular subtraction image generated by registering a fixed mask image with a third frame angiography image, as provided in the embodiments of this specification.

[0066] Figure 9d A schematic diagram of a vascular subtraction image generated by registering the first frame angiography image provided in the embodiments of this specification with its corresponding target matching mask image;

[0067] Figure 9e A schematic diagram of a vascular subtraction image generated by registering the second frame angiography image provided in the embodiments of this specification with its corresponding target matching mask image;

[0068] Figure 9f A schematic diagram of a vascular subtraction image generated by registering the third frame angiography image provided in the embodiments of this specification with its corresponding target matching mask image;

[0069] Figure 10 A flowchart illustrating the process of determining a preliminary set of matching masks provided for the implementation of this specification;

[0070] Figure 11 A schematic flowchart illustrating the vascular subtraction image generation method provided in the embodiments of this specification;

[0071] Figure 12 This is a schematic diagram of a vascular subtraction image generation apparatus provided in the embodiments of this specification. Detailed Implementation

[0072] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0073] An angiography machine is a medical imaging device that assists doctors in examinations or surgeries. It dynamically displays images of the patient's tissues, especially blood vessels, using X-rays. Because blood vessels are not clearly distinguishable from other tissues under X-ray irradiation, a contrast agent is usually injected into the blood vessels to improve their contrast in the image. Then, digital subtraction is used to subtract the image acquired after contrast agent injection (contrast image) from the image acquired without contrast agent injection (masked image), resulting in a subtracted image where only the blood vessels are clearly visible. Digital subtraction relies on the premise that the patient's examination site is in the same position in both the contrast and masked images. This allows the subtracted image to remove the identical background from both images, leaving only the contrast-enhanced blood vessel image. However, in actual examinations, due to the patient's breathing, heartbeat, and movement, the backgrounds of the contrast and masked images are not identical, resulting in significant background residue in the subtracted image, interfering with the display of blood vessels. This is especially true when acquiring chest and abdominal images, where the large range of respiratory movements and potential patient agitation can lead to noticeable background residue in the subtracted image. Although doctors usually ask patients to hold their breath to reduce movement, this is not possible for patients with poor lung function or those undergoing general anesthesia.

[0074] Based on the above analysis, this specification provides a method for generating vascular subtraction images. First, a target angiography image and a preliminary matching mask image that matches the motion patterns of the target angiography image are acquired. Then, the preliminary matching mask image is processed to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image. Finally, a vascular subtraction image corresponding to the target angiography image is generated based on the target matching mask image and the target angiography image. By acquiring a preliminary matching mask image that matches the motion patterns of the target angiography image, the requirements for mask image quality are reduced, better handling situations such as low frame rate acquisition or patient agitation, thereby improving the robustness and reliability of the vascular subtraction image generation method. By processing the preliminary matching mask image that matches the motion patterns of the target angiography image to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image, the similarity of the background between the mask image used to generate the vascular subtraction image and the target angiography image is improved, making the target matching mask image and the target angiography image visually more unified and realistic. Therefore, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image helps to better eliminate background interference in the target angiography image, thereby highlighting the shape and contour of the blood vessels and improving the quality of the vascular subtraction image.

[0075] The vascular subtraction image generation method provided in this specification can be applied to angiography machines. Please refer to [link / reference]. Figure 1aThe angiography subtraction angiography device can include three modules: image acquisition, image processing, and image display. The injection time of the contrast agent is controlled via the console 114, specifically the duration between the start of X-ray emission and the injection of the contrast agent (at least including one respiratory cycle). In the image acquisition module, after the patient is ready, the operator uses a pedal or other control device to trigger the X-ray tube 102, which emits an X-ray beam. During the period between the start of X-ray emission and the injection of the contrast agent, the X-rays pass through the patient's body and are captured by the detector 104. The detector 104 converts the received X-rays into digital signals and records the imaging information. Image reconstruction is performed on the digital signals to generate a masked image sequence. The image memory 106 stores the masked image sequence. The masked image sequence is transmitted to the image processing module in real time. The console 114 in the image processing module controls the mask generation component 108, which performs motion pattern judgment on each frame of the masked image to determine the phase of each frame. A contrast agent is injected, and during the period from injection to the cessation of X-ray emission, a sequence of contrast images is generated by reconstructing the received digital signals. Image memory 106 stores the sequence of contrast images. The sequence of contrast images is transmitted to the image processing module in real time. The mask generation unit 108 in the image processing module determines the phase of each frame of the contrast image by judging its motion pattern. The target contrast image is any frame in the sequence. The mask generation unit 108 determines a preliminary matching mask set in the mask image sequence based on the phase of the target contrast image, and determines a target matching mask image based on the first mask image in the preliminary matching mask set that has the closest phase to the target contrast image, a second mask image (excluding the first mask image), and the target contrast image. The target matching mask image and the target contrast image are transmitted to the subtraction unit 110, and the console 114 controls the subtraction unit 110. The subtraction unit 110 subtracts the target matching mask image from the target contrast image to obtain a vascular subtraction image. The console 114 controls the display 112 to show the angiography image, which is used by doctors to diagnose the patient's condition.

[0076] This specification provides a method for generating vascular subtraction images. Please refer to [link to relevant documentation]. Figure 1b The method for generating vascular subtraction images may include the following steps:

[0077] S110. Obtain the target contrast image and a preliminary matching mask image that matches the motion pattern of the target contrast image.

[0078] The motion pattern can be a certain motion change in a significant tissue structure in a medical image. For example, the motion pattern can be the respiratory phase corresponding to the image, or it can be the positional data of key points in the image.

[0079] Specifically, after the patient is ready, the operator uses a pedal or other control device to trigger the X-ray emitting device, emitting an X-ray beam. The X-rays pass through the patient's body and are captured by a detector. The detector converts the received X-rays into digital signals and records imaging information. The acquired digital signals are transmitted to a computer system for image reconstruction, generating masked images. During the period from the start of X-ray emission to the injection of contrast agent, multiple frames of masked images are generated by reconstructing the received digital signals, and these multiple frames are combined into a masked image sequence. Contrast agent is injected, and during the period from the injection of contrast agent to the cessation of X-ray emission, multiple frames of contrast images are generated by reconstructing the received digital signals, and these multiple frames are combined into a contrast image sequence. The target contrast image is any frame in the contrast image sequence. To obtain a masked image with a background more similar to the target contrast image, it is necessary to match the motion patterns of the target contrast image with the motion patterns of the masked images in the masked image sequence, and masked images that meet the criteria constitute a preliminary matched masked image.

[0080] In some implementations, the determination of image motion patterns can be achieved using an external respiratory gating device. Respiratory gating technology, by fixing a sensor to the patient's chest, can output real-time chest amplitude data, or phase, that changes over time. This data corresponds to the respiratory time at which the images are acquired and can approximate the image's movement. Figure 1c As shown, the horizontal axis represents time, and the vertical axis represents chest amplitude data, i.e., phase. When the patient breathes freely, the diaphragm in the image will show up and down movement, with the upper edge of the diaphragm marked by a red dashed line. When the gating signal amplitude decreases, it indicates that the patient is exhaling, the chest cavity moves downward, and the diaphragm rises; when the gating signal amplitude increases, it indicates that the patient is inhaling, the chest cavity moves upward, and the diaphragm descends. Therefore, by acquiring the respiratory gating signal, the corresponding respiratory phase can be assigned to the masked image and the contrast image in real time to represent the motion pattern of the image.

[0081] S120. Process the preliminary matching mask image to obtain a target matching mask image whose background is similar to that of the target imaging image.

[0082] Specifically, processing the preliminary matching mask image may include noise reduction and / or registration processing, or it may include the fusion processing of multiple preliminary matching mask images as described below. The purpose of processing the preliminary matching mask image is to improve the quality of the matching mask image and eliminate positional differences with the imaging film caused by motion.

[0083] S130. Generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image.

[0084] Among them, "similar background" can mean that the background of the target matching mask image and the background of the target imaging image are more visually consistent or similar.

[0085] Specifically, the preliminary matching mask image is processed to make it consistent with the background of the target angiography image, resulting in a target matching mask image with a background that is nearly identical to that of the target angiography image. Then, subtraction processing is performed on the target matching mask image and the target angiography image to eliminate or reduce vascular signals in the target angiography image, generating a vascular subtraction image corresponding to the target angiography image.

[0086] In the above embodiments, firstly, a target angiography image and a preliminary matching mask image that matches the motion pattern of the target angiography image are acquired. Then, the preliminary matching mask image is processed to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image. Finally, a vascular subtraction image corresponding to the target angiography image is generated based on the target matching mask image and the target angiography image. This embodiment of the specification, by acquiring a preliminary matching mask image that matches the motion pattern of the target angiography image, reduces the requirements for the quality of the mask image and better handles situations such as low frame rate acquisition or patient agitation, thereby improving the robustness and reliability of the vascular subtraction image generation method. By processing the preliminary matching mask image that matches the motion pattern of the target angiography image to obtain a target matching mask image with a background that is nearly identical to that of the target angiography image, the similarity of the background between the mask image used to generate the vascular subtraction image and the target angiography image is improved, making the target matching mask image and the target angiography image more visually unified and realistic. Therefore, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image helps to better eliminate background interference in the target angiography image, thereby highlighting the shape and contour of the blood vessels and improving the quality of the vascular subtraction image.

[0087] This specification provides a method for generating vascular subtraction images. Please refer to [link to relevant documentation]. Figure 2 The method for generating vascular subtraction images may include the following steps:

[0088] S210. Obtain the target contrast image and a preliminary set of matching masks that match the motion patterns of the target contrast image.

[0089] Specifically, during the period from the start of X-ray emission to the injection of contrast agent, multiple mask images are generated by reconstructing the received digital signals, and these multiple mask images are arranged into a mask image sequence. During the period from the injection of contrast agent to the cessation of X-ray emission, multiple contrast images are generated by reconstructing the received digital signals, and these multiple contrast images are arranged into a contrast image sequence. The target contrast image is any frame in the contrast image sequence. To obtain a mask image with a background more similar to the target contrast image, it is necessary to match the motion patterns of the target contrast image with the motion patterns of the mask images in the mask image sequence, and to form a preliminary matching mask set from the multiple mask images that meet the criteria.

[0090] S220. Perform fusion processing on the preliminary matching mask images in the preliminary matching mask set to obtain a target matching mask image whose background is similar to that of the target imaging image.

[0091] S230. Generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image.

[0092] Specifically, the preliminary matching mask images in the preliminary matching mask set are fused to ensure that the fused image retains the detailed information of the preliminary matching mask images and is consistent with the background of the target angiography image, resulting in a target matching mask image with a background that is similar to that of the target angiography image. Then, subtraction processing is performed on the target matching mask image and the target angiography image to eliminate or reduce the vascular signal in the target angiography image, generating a vascular subtraction image corresponding to the target angiography image.

[0093] In the above implementation, firstly, a target angiography image and a preliminary matching mask set that matches the motion patterns of the target angiography image are acquired. Then, the preliminary matching mask images in the preliminary matching mask set are fused to obtain a target matching mask image with a background that is similar to that of the target angiography image. Finally, a vascular subtraction image corresponding to the target angiography image is generated based on the target matching mask image and the target angiography image. By acquiring a preliminary matching mask set that matches the motion patterns of the target angiography image, the requirements for mask image quality are reduced, better handling situations such as low frame rate acquisition or patient agitation, thereby improving the robustness and reliability of the vascular subtraction image generation method. By fusing the preliminary matching mask images in the preliminary matching mask set that matches the motion patterns of the target angiography image to obtain a target matching mask image with a background that is similar to that of the target angiography image, the similarity of the background between the mask image used to generate the vascular subtraction image and the target angiography image is improved, making the target matching mask image and the target angiography image more visually unified and realistic. Therefore, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image helps to better eliminate background interference in the target angiography image, thereby highlighting the shape and contour of the blood vessels and improving the quality of the vascular subtraction image.

[0094] In some implementations, please refer to Figure 3a The initial set of matching masks is determined in the following way:

[0095] S310. Obtain the mask image sequence.

[0096] Specifically, once the patient is ready, the operator uses a pedal or other control device to trigger the X-ray emitting device, emitting an X-ray beam. The X-rays pass through the patient's body and are captured by a detector. The detector converts the received X-rays into digital signals and records imaging information. The acquired digital signals are transmitted to a computer system for image reconstruction, generating masked images. During the period from the start of X-ray emission to the injection of contrast agent, multiple frames of masked images are generated through reconstruction processing of the received digital signals, and these multiple frames are combined into a masked image sequence.

[0097] S320. Determine the target phase stage of the target contrast image based on the phase of the target contrast image and the phase of the specified set of contrast images.

[0098] Specifically, at least a portion of the specified set of contrast images is located before the target contrast image, or the specified set of contrast images is located after the target contrast image. The target phase phase can be the phase state or stage of the target contrast image, such as an ascending phase or a descending phase.

[0099] Specifically, rules or standards for determining a specified set of contrast images can be pre-defined based on phase differences, time information, or other relevant factors. The specified set of contrast images is determined according to these pre-defined rules or standards, with the phase of the target contrast image as a reference. At least some of the specified contrast images are located before or after the target contrast image. The phase of the specified contrast images is determined. Then, based on the positional relationship between the specified contrast images and the target contrast image, and the magnitude relationship between the phase of each frame in the specified contrast images and the phase of the target contrast image, it is determined whether the target phase of the target contrast image is in an ascending or descending phase.

[0100] For example, please refer to Figure 3b In the figure, the discrete black dots on the curve represent the acquired masked or angiographic images. c For contrast imaging, use contrast imaging image I c The first two frames of contrast-enhanced images are identified as the designated contrast-enhanced image set. Based on the phase of the target contrast-enhanced image and the phase of each frame in the designated contrast-enhanced image set, the target phase of the target contrast-enhanced image is determined to be the rising phase. The number of frames in the designated contrast-enhanced image set can be determined by increasing or decreasing the acquisition frame rate. A higher acquisition frame rate results in a larger number of frames in the designated contrast-enhanced image set.

[0101] S330. Determine a preliminary matching mask set in the mask image sequence based on the target phase stage and the motion law data of the target imaging image.

[0102] Specifically, the motion pattern data of the target contrast image is matched with the motion pattern data of the masked image sequence within the target phase stage. Based on the matching results, multiple masked images with backgrounds similar to the target contrast image are selected to form a preliminary matching mask set.

[0103] In the above embodiments, a mask image sequence is obtained, the target phase stage of the target contrast image is determined based on the phase of the target contrast image and the phase of the specified contrast image set, and a preliminary matching mask set is determined in the mask image sequence based on the target phase stage and the motion law data of the target contrast image, so as to provide a data basis for determining the target matching mask image with a background that is similar to that of the target contrast image.

[0104] In some implementations, a preliminary matching mask image is determined by: acquiring a mask image sequence; determining the target phase stage of the target contrast image based on the phase of the target contrast image and the phase of a specified set of contrast images; and determining a preliminary matching mask image in the mask image sequence based on the target phase stage and motion law data of the target contrast image.

[0105] Specifically, during the period from the start of X-ray emission to the injection of contrast agent, multiple mask images are generated by reconstructing the received digital signals, and these multiple mask images are arranged into a mask image sequence. A specified set of contrast images is determined according to pre-set rules or standards, using the phase of the target contrast image as a reference. At least some of the specified contrast images are located before or after the target contrast image. The phase of the specified contrast images is determined. Then, based on the positional relationship between the specified contrast images and the target contrast image, and the magnitude relationship between the phase of each frame in the specified contrast images and the phase of the target contrast image, it is determined whether the target phase stage of the target contrast image is in an ascending or descending phase. The motion data of the target contrast image is matched with the motion data of the mask images in the mask image sequence within the target phase stage. Based on the matching results, a mask image with a background similar to the target contrast image is selected as the initial matching mask image.

[0106] In the above embodiments, a mask image sequence is obtained, the target phase stage of the target contrast image is determined based on the phase of the target contrast image and the phase of the specified contrast image set, and a preliminary matching mask image is determined in the mask image sequence based on the target phase stage and the motion law data of the target contrast image, so as to provide a data basis for determining the target matching mask image with a background that is similar to that of the target contrast image.

[0107] In some implementations, please refer to Figure 4a The masked image sequence corresponds to multiple motion cycles. Determining a preliminary matching mask set in the masked image sequence based on the target phase stage and motion pattern data of the target imaging image may include the following steps:

[0108] S410. Determine multiple masked images located within the target phase phase in each motion cycle of the masked image sequence.

[0109] S420. In each motion cycle, among the multi-frame mask images located in the target phase phase, determine the set of reference mask images whose similar data to the specified angiographic image set meets the preset similarity screening conditions.

[0110] S430. In the masked image sequence, determine multiple masked images that are located within the motion cycle of the reference masked image set and match the motion pattern data of the target imaging image, and use them as a preliminary matching masked image set.

[0111] The preset similarity filtering criteria can be a set of conditions or standards set when determining a set of reference mask images similar to a specified set of contrast images. For example, the preset similarity filtering criteria can be a similarity measure based on motion pattern data. The motion cycle can be the periodic change of human movement, such as including the respiratory cycle or the heartbeat cycle.

[0112] Specifically, after determining the target phase stage, the masked image sequence is analyzed to determine each motion cycle in the masked image sequence and the target phase stage within each motion cycle. All masked images within the target phase stage of each motion cycle can be considered as multiple frames of masked images within that target phase stage in each motion cycle. Alternatively, the masked image within the target phase stage of each motion cycle that most closely matches the motion pattern data of a specified contrast image set and has the same frame number as the specified contrast image set can be considered as multiple frames of masked images within the target phase stage in each motion cycle.

[0113] The similarity data between multiple masked images located within the target phase phase in each motion cycle and a specified set of contrast images is calculated. By comparing the similarity data, multiple masked images located within the target phase phase in a certain motion cycle whose similarity data meets the preset similarity screening criteria are identified as a reference masked image set. Then, in the masked image sequence, the motion law data corresponding to each masked image within the motion cycle of the reference masked image set is matched with the motion law data of the target contrast image. The masked image whose motion law data is closest to that of the target contrast image is selected as the preliminary matching masked image. Then, the closest masked image among its adjacent masked images is selected as the preliminary matching masked image. The resulting multiple preliminary matching masked images constitute a preliminary matching masked set.

[0114] For example, please refer to Figure 4b P1 is a set of reference mask images. In the mask image sequence, the mask image I that is located within the motion cycle of the reference mask image set P1 and whose motion pattern data is closest to that of the target contrast image is identified. 1-1 and masked image I 1-1 Adjacent masked images I 1-2 This forms a preliminary set of matching masks.

[0115] In the above embodiments, multiple mask images located within the target phase phase in each motion cycle are determined in the mask image sequence. Among the multiple mask images located within the target phase phase in each motion cycle, a set of reference mask images whose similar data with the specified contrast image set meets the preset similarity screening conditions is determined. In the mask image sequence, multiple mask images located within the motion cycle of the reference mask image set and matching the motion pattern data of the target contrast image are determined as a preliminary matching mask set, providing a data basis for determining the target matching mask image whose background tends to be the same as the target contrast image.

[0116] In some implementations, the mask image sequence corresponds to multiple motion cycles. Determining a preliminary matching mask image in the mask image sequence based on the motion pattern data of the target phase stage and the target contrast image may include: determining multiple mask images located within the target phase stage in each motion cycle of the mask image sequence; determining a set of reference mask images whose similarity data with a specified contrast image set satisfies preset similarity screening conditions among the multiple mask images located within the target phase stage in each motion cycle; and determining, within the mask image sequence, a mask image located within the motion cycle of the reference mask image set and matching the motion pattern data of the target contrast image, as the preliminary matching mask image.

[0117] Specifically, the masked image sequence is analyzed to determine each motion cycle and the target phase phase within each motion cycle. All masked images located within the target phase phase of each motion cycle can be considered as multi-frame masked images within that phase. Alternatively, the masked image located within the target phase phase of each motion cycle that most closely matches the motion pattern data of a specified contrast image set and has the same frame number as the specified contrast image set can be considered as multi-frame masked images within the target phase phase of each motion cycle. Similarity data between the multi-frame masked images and the specified contrast image set is calculated. By comparing the similarity data, multi-frame masked images within the target phase phase of a certain motion cycle whose similarity data meets preset similarity screening conditions are identified as a reference masked image set. Then, in the masked image sequence, the motion pattern data corresponding to each masked image within the motion cycle of the reference masked image set is matched with the motion pattern data of the target contrast image. The masked image whose motion pattern data is closest to that of the target contrast image is selected as the initial matching masked image.

[0118] In the above embodiments, multiple mask images located within the target phase phase in each motion cycle are determined in the mask image sequence. Among the multiple mask images located within the target phase phase in each motion cycle, a set of reference mask images whose similar data with the specified contrast image set meets the preset similarity screening conditions is determined. In the mask image sequence, a mask image located within the motion cycle of the reference mask image set and matching the motion pattern data of the target contrast image is determined as a preliminary matching mask image, providing a data basis for determining a target matching mask image whose background tends to be the same as the target contrast image.

[0119] In some implementations, please refer to Figure 5a In each motion cycle, among multiple masked images located within the target phase phase, determining a set of reference masked images whose similar data to a specified set of contrast images meets preset similarity screening criteria may include the following steps:

[0120] S510. Based on the motion pattern data of the specified set of contrast images, match the multi-frame mask images located in the target phase phase in each motion cycle to obtain the set of mask images to be screened in each motion cycle.

[0121] S520. In the set of mask images to be screened for each of the multiple motion cycles, determine the set of reference mask images that are most similar to the specified set of contrast images.

[0122] Specifically, after determining the target phase, the motion pattern data corresponding to the mask images located within the target phase in each motion cycle is matched with the motion pattern data of a specified angiographic image set. This determines the multi-frame mask images within the target phase in each motion cycle that are closest to the motion pattern data of the specified angiographic image set, which are then used as the mask set to be screened. The similarity between the mask set to be screened and the specified angiographic image set in each motion cycle is calculated. The similarity between the mask set to be screened and the specified angiographic image set in each motion cycle is compared. Among the mask sets to be screened in each of the multiple motion cycles, the mask set most similar to the specified angiographic image set is selected as the reference mask set most similar to the specified angiographic image set. The number of frames of angiographic images in the specified angiographic image set can be determined by increasing or decreasing the acquisition frame rate. A higher acquisition frame rate results in a larger number of frames of angiographic images in the specified angiographic image set.

[0123] For example, the target phase is determined to be the rising phase. See also Figure 5b P c For a specified set of contrast images, determine the mask images corresponding to the specified set of contrast images P within the rising phase of each motion cycle. cThe multi-frame mask images with the closest phase are selected as the mask set P1 and the mask set P2 to be filtered. The mask set P1 and the mask set P2 are compared with the specified angiographic image set P using the formula shown below. c Similarity:

[0124]

[0125] in, For a specified set of contrast images P c The i-th frame mask image, For the set of masking elements to be filtered, P n The i-th frame mask image. M is a specified set of contrast images containing M frames. S n For the set of contrast images P c With the set of masking elements to be filtered P n The similarity between them.

[0126] S n A smaller value indicates a better match to the motion state, allowing for the prioritization of filtering out interference from sudden movements. The sets of mask images to be screened, P1 and P2, for each motion cycle are compared with the specified set of contrast images, P... c The similarity between the images can be used to determine that P1 is the set of reference mask images that is most similar to the specified set of contrast images.

[0127] In the above embodiments, based on the motion pattern data of the specified angiographic image set, matching is performed on multiple mask images located in the target phase phase in each motion cycle to obtain a mask set to be screened in each motion cycle. Among the mask sets to be screened in each of the multiple motion cycles, a set of reference mask images most similar to the specified angiographic image set is determined, providing a data basis for subsequently determining the first mask image, the second mask image, and thus determining the target matching mask image whose background tends to be the same as the target angiographic image.

[0128] In some implementations, please refer to Figure 6 Multiple preliminary matching mask images are generated, forming a preliminary matching mask set. Determining the preliminary matching mask image from the mask image sequence based on the target phase stage and motion pattern data of the target imaging image may include the following steps:

[0129] S610. Determine the multiple masked images located within the target phase phase in the masked image sequence.

[0130] S620. Based on the motion law data of the target imaging image, match the multiple mask images located in the target phase phase to obtain a preliminary matching mask set.

[0131] Specifically, the masked image sequence is analyzed, and all masked images located within the target phase change phase are identified as multi-frame masked images within the target phase phase, denoted as the same-phase mask set. For example, the multi-frame masked images within the target phase phase could be all masked images in the rising phase or all masked images in the falling phase. Then, the similarity data between the motion law data of the target contrast image and the motion law data of each frame masked image in the same-phase mask set is calculated. Finally, by comparing the similarity data between the target contrast image and each frame masked image in the same-phase mask set, matching is achieved based on the motion law data of the target contrast image within the target phase phase of the multi-frame masked images. The masked image corresponding to the highest similarity data is used as the preliminary matching masked image. Then, the masked image corresponding to the highest similarity data among the adjacent masked images is used as the preliminary matching masked image. The determined multi-frame preliminary matching masked images constitute the preliminary matching mask set.

[0132] In the above embodiments, multiple mask images located within the target phase phase are determined in the mask image sequence. Based on the motion law data of the target angiography image, matching is performed on the multiple mask images located within the target phase phase to obtain a preliminary matching mask set, which provides a data basis for determining the target matching mask image with a background that is similar to that of the target angiography image.

[0133] In some implementations, there are multiple preliminary matching mask images, which together form a preliminary matching mask set. The target matching mask image is obtained by fusing the preliminary matching mask images from the preliminary matching mask set. Processing the preliminary matching mask images to obtain a target matching mask image with a background that closely resembles the target imaging image may include: performing keypoint localization and pixel value calculation based on the preliminary matching mask images from the preliminary matching mask set to obtain the target matching mask image.

[0134] Keypoints can represent the salient organizational structure of an image, such as corners, edges, textures, and other prominent locations.

[0135] Specifically, keypoint detection is performed on the preliminary matching mask images in the preliminary matching mask set to determine the locations of significant changes or features in the preliminary matching mask images, thereby achieving keypoint localization. After determining the keypoints, based on the pixel values ​​and positions of the keypoints in the preliminary matching mask images, the pixel values ​​at the corresponding keypoint positions in the target matching mask image can be calculated, thus obtaining the target matching mask image.

[0136] In the above embodiments, key point localization and pixel value calculation are performed based on the preliminary matching mask images in the preliminary matching mask set to obtain the target matching mask image. By obtaining the target matching mask image with a background that is similar to that of the target angiography image, a data basis is provided for eliminating background interference in the target angiography image, highlighting the shape and contour of blood vessels, and thus improving the quality of the vascular subtraction image.

[0137] In some implementations, please refer to Figure 7 The initial matching mask set includes a first mask image and a second mask image. Based on the initial matching mask images in the initial matching mask set, keypoint localization and pixel value calculation are performed to obtain the target matching mask image, which may include the following steps:

[0138] S710. Determine the first position data corresponding to the key points in the first masking image and the second position data corresponding to the key points in the second masking image.

[0139] In this process, key points in the first masked image are matched with key points in the second masked image.

[0140] Specifically, gradient calculation is performed on the first mask image to obtain its corresponding gradient map. Then, gradient points that meet specific conditions (such as local maxima) are selected from the gradient map and designated as keypoints to locate them, thus obtaining the first position data corresponding to the keypoints in the first mask image. Then, based on predefined matching rules (such as distance between keypoints, feature similarity, etc.), the first position data is used as a reference to determine the second position data corresponding to the keypoints in the second mask image.

[0141] For example, a global gradient is calculated on the first mask image to obtain a gradient map. Then, a thresholding method is used to divide the gradient map into multiple regions, and the point with the largest gradient within each region is selected as a keypoint. Based on the determined keypoints, the first position data corresponding to the keypoints is determined in the first mask image.

[0142] S720. Using the first phase data of the first mask image, the second phase data of the second mask image, the third phase data of the target imaging image, the first position data and the second position data, interpolation processing is performed to obtain the target position data of the key points in the target matching mask image.

[0143] Specifically, an appropriate interpolation method (such as bilinear interpolation) is selected based on the actual situation and specific needs. Using the selected interpolation method, interpolation is performed on the first phase data of the first mask image, the second phase data of the second mask image, the third phase data of the target imaging image, the first position data, and the second position data. Data for unknown positions is inferred from the known position data to obtain more continuous and smooth position data, thus obtaining the target position data of key points in the target matching mask image.

[0144] For example, the formula for obtaining the target location data of key points in the target matching mask image using bilinear interpolation is shown below:

[0145]

[0146] Among them, A c For the third phase data of the target imaging image, D 1-1 For the first position data, D is the first phase data of the first mask image. 1-2 For the second position data, This refers to the second phase data of the second mask image. (D) m Match the target location data of key points in the masked image to the target.

[0147] S730. Based on the pixel data of the first mask image, the pixel data of the second mask image, the first position data, and the second position data, interpolation processing is performed to obtain the pixel values ​​of key points in the target matching mask image.

[0148] Specifically, an appropriate interpolation method (such as bilinear interpolation or bicubic interpolation) is selected based on the actual situation and specific needs to estimate the pixel values ​​of unknown pixels in the interpolation region. The location data of key points on the first mask image are determined based on the first location data, and the location data of key points on the second mask image are determined based on the second location data. Using the selected interpolation method, pixel values ​​are interpolated for the pixel data at the key points in both the first and second mask images to obtain more continuous and smoother pixel values ​​for the target location data of key points in the target matching mask image. The above steps are repeated until pixel values ​​for all key points have been interpolated.

[0149] S740. Obtain a target matching mask image based on the target location data and pixel values ​​of key points in the target matching mask image.

[0150] Specifically, using the target location data and pixel values ​​of key points in the target matching mask image, a portion of the target matching mask image can be generated. Gradient points in the first mask image, excluding key points, are designated as non-key points. Then, based on the location data corresponding to the non-key points in the first mask image, matching is performed to determine the location data corresponding to the non-key points in the second mask image. The remaining portion of the target matching mask image can be generated by fusing the location data and pixel values ​​of the non-key points. Finally, the portion of the target matching mask image obtained above is merged with the remaining portion of the target matching mask image to generate the complete target matching mask image.

[0151] For example, in the first mask image, the first mask image is divided into several triangular regions based on its key points. For each non-key point within a triangular region, its position is determined based on the key point positions corresponding to the three vertices of that region and the distance relationships between the three vertices. Based on the distance relationships between the key points in the first and second mask images, a weighted interpolation method is used; that is, the closer a point is to a vertex of the triangle, the greater its weight in determining its position. This determines the position and pixel value of the non-key points in the target matching mask image.

[0152] In the above embodiments, first position data corresponding to key points in the first mask image and second position data corresponding to key points in the second mask image are determined. Interpolation processing is performed using first phase data of the first mask image, second phase data of the second mask image, third phase data of the target angiography image, first position data, and second position data to obtain target position data of key points in the target matching mask image. Interpolation processing is performed based on pixel data of the first mask image, pixel data of the second mask image, first position data, and second position data to obtain pixel values ​​of key points in the target matching mask image. Based on the target position data and pixel values ​​of key points in the target matching mask image, a target matching mask image is obtained. By obtaining a target matching mask image with a background that is similar to that of the target angiography image, a data foundation is provided for eliminating background interference in the target angiography image, highlighting the morphology and contour of blood vessels, and thus improving the quality of the vascular subtraction image.

[0153] In some implementations, please refer to Figure 8a The initial matching mask set has two frames of images. Before determining the first position data corresponding to the key points in the first mask image, the method may further include the following steps:

[0154] S810. In the preliminary matching mask set, determine the preliminary matching mask image that is closest to the motion law data of the target imaging image, and use it as the first mask image.

[0155] S820. Use the preliminary matching mask images other than the first mask image in the preliminary matching mask set as the second mask image.

[0156] The second position data is determined on the second mask image based on the first position data using template matching.

[0157] Specifically, the motion pattern data of the target contrast image is compared with the motion pattern data of each frame of the preliminary matching mask image in the preliminary matching mask set. Based on the comparison results, the preliminary matching mask image that is closest to the motion pattern data of the target contrast image is selected as the first mask image. Then, the selected first mask image is excluded from the preliminary matching mask set, and the remaining preliminary matching mask images in the preliminary matching mask set are used as the second mask images. This ensures that the motion patterns of the mask images complement those of the target contrast image, thereby improving the registration effect and stability.

[0158] For example, the initial matching mask set includes I 1-1 and I 1-2 Please see. Figure 8b The phase of the target contrast image is compared with the phase of each frame of the preliminary matching mask image in the preliminary matching mask set, and the preliminary matching mask image I with the closest phase to the target contrast image is selected. 1-1 As the first mask image, the initial matching mask set excluding the first mask image I is used. 1-1 Preliminary matching mask image I 1-2 As a second masking image.

[0159] In the above embodiments, the preliminary matching mask image that is closest to the motion pattern data of the target angiography image is determined from the preliminary matching mask set and used as the first mask image. The preliminary matching mask images other than the first mask image in the preliminary matching mask set are used as the second mask images, so as to provide a data basis for the subsequent generation of the vascular subtraction image corresponding to the target angiography image.

[0160] In some implementations, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image may include: registering the target matching mask image and the target angiography image to obtain a registered mask image; and subtracting the registered mask image from the target angiography image to obtain the vascular subtraction image.

[0161] Registration can be the process of aligning or calibrating the target matching mask image with the target imaging image so that they are spatially consistent or accurately overlapped.

[0162] Specifically, an appropriate registration method (such as two-dimensional elastic registration or registration using deep learning methods) is selected based on the specific circumstances. Then, the target matching mask image and the target angiography image are registered using the selected registration method to obtain a registered mask image. This ensures that the background between the registered mask image and the target angiography image is more similar than that between the target matching mask image and the target angiography image. Therefore, to more clearly present the vascular structure, the pixel value at the corresponding position in the target angiography image is subtracted from the pixel value at the corresponding position in the registered mask image pixel by pixel. This ensures that background interference is effectively removed at each pixel level, allowing the vascular structure to be displayed more clearly, resulting in a vascular subtraction image.

[0163] In the above embodiments, registration is performed based on the target matching mask image and the target angiography image to obtain a registered mask image. The registered mask image is then subtracted from the target angiography image to obtain a vascular subtraction image. By using a target matching mask image with a background similar to the target angiography image for registration, the similarity between the registered mask image and the target angiography image is improved. Then, the background interference in the target angiography image is eliminated using the target angiography image, further highlighting the morphology and contour of the blood vessels, thereby improving the quality of the vascular subtraction image.

[0164] In some implementations, generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image may include: subtracting the target matching mask image from the target angiography image to obtain the vascular subtraction image.

[0165] Specifically, since target matching mask images usually contain background structures unrelated to blood vessels, in order to present the blood vessel structure more clearly, the pixel value of the corresponding position in the target angiography image is subtracted from the pixel value of the same position in the target matching mask image pixel by pixel. This ensures that background interference is effectively removed at each pixel level, so that the blood vessel structure can be displayed more clearly, resulting in a blood vessel subtraction image.

[0166] For example, compared to related technologies that use a fixed mask image and register all angiography images with it to generate a vascular subtraction image corresponding to each angiography image, for example, please refer to Figure 9a , Figure 9a This is a subtraction angiography image generated by registering a mask image with the first angiography image. Please refer to [link to relevant documentation]. Figure 9b , Figure 9b This is a subtraction angiography image generated by registering a mask image with a second angiographic image. (See also...) Figure 9c , Figure 9cThis specification describes a method for generating a vascular subtraction image by registering a mask image with a third frame of angiography. In this embodiment, the target-matching mask image corresponding to each frame of angiography is used for registration to generate a higher-quality vascular subtraction image. For example, please refer to [link to relevant documentation]. Figure 9d , Figure 9d The resulting vascular subtraction image is generated by registering the first angiographic image with its corresponding target-matching mask image. (See also...) Figure 9e , Figure 9e The resulting subtraction angiography image is generated by registering the second angiography image with its corresponding target-matching mask image. (See also...) Figure 9f , Figure 9f The vascular subtraction image is generated by registering the third frame angiography image with its corresponding target matching mask image.

[0167] In the above embodiments, a vascular subtraction image is obtained by subtracting a target matching mask image from a target angiography image. By using a target matching mask image with a background that is similar to that of the target angiography image to eliminate background interference in the target angiography image, the shape and contour of the blood vessels are highlighted more, thereby improving the quality of the vascular subtraction image.

[0168] In some implementations, please refer to Figure 10 The initial matching mask images consist of multiple images, which together form an initial matching mask set. The initial matching mask set is determined using the following method:

[0169] S1010, Obtain the mask image sequence.

[0170] Specifically, once the patient is ready, the operator uses a pedal or other control device to trigger the X-ray emitting device, emitting an X-ray beam. The X-rays pass through the patient's body and are captured by a detector. The detector converts the received X-rays into digital signals and records imaging information. The acquired digital signals are transmitted to a computer system for image reconstruction, generating masked images. During the period from the start of X-ray emission to the injection of contrast agent, multiple frames of masked images are generated through reconstruction processing of the received digital signals, and these multiple frames are combined into a masked image sequence.

[0171] S1020, Determine the third position data of key points of the target imaging image and the fourth position data of key points of each frame of the masked image in the masked image sequence.

[0172] Specifically, gradient calculation is performed on the first frame of the masked image sequence to obtain the first gradient map corresponding to the first frame. Then, gradient points that meet specific conditions (such as local maxima) are selected from the first gradient map and used as keypoints to locate them, thus obtaining the positional data of the keypoints in the first frame. Next, based on set matching rules (such as distance between keypoints, feature similarity, etc.), the positional data of the keypoints in the first frame is used as a reference to determine the positional data of the keypoints in the second frame. Through iterative loops, in each iteration, the positional data of the keypoints in the current frame is determined based on the positional data of the keypoints in the previous frame, thereby determining the fourth positional data of the keypoints in each frame of the masked image sequence. Finally, based on the matching rules, the last frame of the masked image sequence is used as a reference to determine the positional data of the keypoints in the first frame of the contrast image. The process is repeated iteratively. In each iteration, the keypoint location data in the current frame of the contrast-enhanced image is determined based on the keypoint location data in the previous frame, thus determining the keypoint location data for each frame in the contrast-enhanced image sequence. Based on the keypoint location data for each frame in the contrast-enhanced image sequence, the third location data of the keypoints in the target contrast-enhanced image is determined.

[0173] For example, a global gradient is calculated on the first frame of the masked image sequence to obtain a gradient map. Then, a thresholding method is used to divide the gradient map into multiple regions, and the point with the largest gradient in each region is selected as a keypoint. This ensures that the number of keypoints is neither too large nor too sparse. Subsequently, following inter-frame tracking and template matching methods, the gradient keypoints in the first frame of the masked image are tracked in each subsequent frame of the masked image sequence to determine the fourth position data of the keypoints in each frame of the masked image sequence. The fourth position data of the keypoints in each frame of the masked image sequence is determined using the following formula:

[0174] D n = Template Matching (M n -M n-1 )+D n-1

[0175] Among them, M n Let M be the nth frame mask image in the mask image sequence. n-1 This is the (n-1)th frame of the masked image sequence. (D) n-1 D is the fourth position data of key point K1 in the (n-1)th frame mask image. n This is the fourth position data of key point K1 in the nth frame mask image.

[0176] Similarly, the third location data of key points in the target imaging image are determined using inter-frame tracking and template matching methods. It should be noted that by employing inter-frame tracking and template matching, the locations of key points can be determined more accurately, avoiding getting trapped in local minima and significantly accelerating the displacement calculation.

[0177] S1030. Based on the third position data and the fourth position data, determine the image similarity data between each frame of the mask image and the target contrast image in the mask image sequence.

[0178] S1040. Determine a preliminary matching mask set in the mask image sequence based on image similarity data.

[0179] Among them, image similarity data can represent the degree of background similarity between each frame of the masked image and the target imaging image in the masked image sequence.

[0180] Specifically, using similarity calculation methods (such as cosine similarity, dot product, Euclidean distance, etc.), the similarity between the third position data of key points in the target contrast image and the fourth position data of key points in each frame of the masked image sequence is calculated to determine the image similarity data between each frame of the masked image sequence and the target contrast image. The masked image corresponding to the image similarity data with the highest similarity is used as the initial matching masked image. Then, the masked images corresponding to the image similarity data with the highest similarity among the adjacent masked images of the masked image corresponding to the highest similarity are used as the initial matching masked images. Based on the determined multiple frames of initial matching masked images, an initial matching mask set is constructed.

[0181] For example, the Euclidean distance method is used to determine the sum of Euclidean distances between each frame of the masked image sequence and the target contrast image. The masked image corresponding to the minimum Euclidean distance sum is used as the initial matching masked image. Then, the masked images corresponding to the minimum Euclidean distance sum among the neighboring masked images of the minimum Euclidean distance sum are used as the initial matching masked images. The initial matching masked image set is constructed based on the determined multiple frames of initial matching masked images.

[0182] In the above embodiments, a mask image sequence is acquired, the third position data of the key points of the target angiography image and the fourth position data of the key points of each frame of the mask image in the mask image sequence are determined, and the image similarity data between each frame of the mask image in the mask image sequence and the target angiography image is determined based on the image similarity data, and a preliminary matching mask set is determined in the mask image sequence based on the image similarity data, so as to provide a data basis for improving the quality of the angiography image in the future.

[0183] This specification also provides a method for generating vascular subtraction images, where the mask image sequence corresponds to multiple motion cycles, including respiratory cycles or cardiac cycles. The initial matching mask set has two frames of images, including a first mask image and a second mask image. For example, please refer to... Figure 11 The method for generating vascular subtraction images may include the following steps:

[0184] S1102, Obtain the target contrast image and obtain the mask image sequence.

[0185] S1104. Determine the target phase stage of the target contrast image based on the phase of the target contrast image and the phase of the specified contrast image set.

[0186] Wherein, at least a portion of the specified set of contrast images is located before the target contrast image, or the specified set of contrast images is located after the target contrast image.

[0187] S1106. Determine multiple masked images located within the target phase phase in each motion cycle of the masked image sequence.

[0188] S1108. Based on the motion pattern data of the specified set of contrast images, match the multiple mask images located in the target phase phase in each motion cycle to obtain the set of mask images to be screened in each motion cycle.

[0189] S1110. In the set of mask images to be screened for each of the multiple motion cycles, determine the set of reference mask images that are most similar to the specified set of contrast images.

[0190] S1112. In the masked image sequence, determine multiple masked images that are located within the motion cycle of the reference masked image set and match the motion pattern data of the target imaging image, and use them as a preliminary matching masked set.

[0191] S1114. In the preliminary matching mask set, determine the preliminary matching mask image that is closest to the motion law data of the target imaging image, and use it as the first mask image.

[0192] S1116. Take the preliminary matching mask images other than the first mask image in the preliminary matching mask set as the second mask images.

[0193] The second position data is determined on the second mask image based on the first position data using template matching.

[0194] S1118. Determine the first position data corresponding to the key points in the first masking image and the second position data corresponding to the key points in the second masking image.

[0195] In this process, key points in the first masked image are matched with key points in the second masked image.

[0196] S1120. Using the first phase data of the first mask image, the second phase data of the second mask image, the third phase data of the target imaging image, the first position data and the second position data, interpolation processing is performed to obtain the target position data of the key points in the target matching mask image.

[0197] S1122. Based on the pixel data of the first mask image, the pixel data of the second mask image, the first position data, and the second position data, interpolation processing is performed to obtain the pixel values ​​of key points in the target matching mask image.

[0198] S1124. Obtain a target matching mask image based on the target location data and pixel values ​​of key points in the target matching mask image.

[0199] S1126. Subtract the target matching mask image from the target angiography image to obtain the blood vessel subtraction image.

[0200] This specification provides a vascular subtraction image generation device 1200. Please refer to [link to documentation]. Figure 12 The vascular subtraction image generation device 1200 includes: an image acquisition module 1210, an image processing module 1220, and a subtraction image generation module 1230.

[0201] The image acquisition module 1210 is used to acquire a target contrast image and a preliminary matching mask image that matches the motion pattern of the target contrast image;

[0202] Image processing module 1220 is used to process the preliminary matching mask image to obtain a target matching mask image whose background is similar to that of the target imaging image;

[0203] The subtraction image generation module 1230 is used to generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image.

[0204] For a detailed description of the angiography image generation device, please refer to the description of the angiography image generation method above, which will not be repeated here.

[0205] In some embodiments, a medical imaging device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method steps described above.

[0206] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in any of the above embodiments.

[0207] One embodiment of this specification provides a computer program product including instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0208] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

Claims

1. A method for generating vascular subtraction images, characterized in that, The method includes: Acquire the target contrast image and a preliminary matching mask image that matches the motion pattern of the target contrast image; The preliminary matching mask image is processed to obtain a target matching mask image whose background is similar to that of the target imaging image. The processing of the preliminary matching mask image includes noise reduction and / or registration processing, or includes fusion processing of multiple preliminary matching mask images. Generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image; The preliminary matching mask images are multiple images, and these multiple preliminary matching mask images constitute a preliminary matching mask set. The target matching mask image is obtained by fusing the preliminary matching mask images in the preliminary matching mask set. The step of processing the preliminary matching mask images to obtain a target matching mask image with a background that is similar to the target imaging image includes: performing key point localization and pixel value calculation based on the preliminary matching mask images in the preliminary matching mask set to obtain the target matching mask image. The preliminary matching mask set includes a first mask image and a second mask image; the step of locating key points and calculating pixel values ​​based on the preliminary matching mask images in the preliminary matching mask set to obtain the target matching mask image includes: Determine the first position data corresponding to key points in the first masking image and the second position data corresponding to key points in the second masking image; wherein, the key points in the first masking image match the key points in the second masking image; Interpolation processing is performed using the first phase data of the first mask image, the second phase data of the second mask image, the third phase data of the target imaging image, the first position data, and the second position data to obtain the target position data of key points in the target matching mask image; Interpolation processing is performed on the pixel data of the first mask image, the pixel data of the second mask image, the first position data, and the second position data to obtain the pixel values ​​of key points in the target matching mask image; The target matching mask image is obtained based on the target location data and pixel values ​​of key points in the target matching mask image.

2. The method according to claim 1, characterized in that, The preliminary matching mask image is determined using the following method: Obtain the mask image sequence; The target phase stage of the target contrast image is determined based on the phase of the target contrast image and the phase of a specified set of contrast images; wherein at least a portion of the specified set of contrast images is located before the target contrast image, or the specified set of contrast images is located after the target contrast image; The preliminary matching mask image is determined in the mask image sequence based on the target phase stage and the motion law data of the target imaging image.

3. The method according to claim 2, characterized in that, The mask image sequence corresponds to multiple motion cycles; determining the preliminary matching mask image in the mask image sequence based on the target phase stage and the motion law data of the target imaging image includes: In the masked image sequence, determine multiple frames of masked images located within the target phase phase of each motion cycle; In each motion cycle, among the multi-frame masked images located within the target phase phase, a set of reference masked images whose similar data to the specified angiographic image set satisfies preset similarity screening conditions is determined; In the masked image sequence, a masked image that is located within the motion cycle of the reference masked image set and matches the motion pattern data of the target imaging image is identified as the preliminary matching masked image.

4. The method according to claim 3, characterized in that, The process of determining a set of reference mask images that satisfy preset similarity screening conditions with the specified angiographic image set from among the multi-frame mask images located within the target phase phase in each motion cycle includes: Based on the motion pattern data of the specified angiographic image set, match the multi-frame mask images located in the target phase phase in each motion cycle to obtain the mask set to be screened in each motion cycle; In each of the multiple motion cycles, the set of reference mask images that is most similar to the specified set of contrast images is determined.

5. The method according to claim 2, characterized in that, The preliminary matching mask images are multiple images, and the multiple preliminary matching mask images constitute a preliminary matching mask set. The step of determining the preliminary matching mask images from the mask image sequence based on the target phase stage and the motion law data of the target imaging image includes: Determine multiple frames of masked images located within the target phase phase in the masked image sequence; Based on the motion pattern data of the target imaging image, matching is performed in multiple mask images located within the target phase phase to obtain the preliminary matching mask set.

6. The method according to claim 1, characterized in that, The preliminary matching mask set has two frames of images; before determining the first position data corresponding to the key points in the first mask image, the method further includes: In the preliminary matching mask set, the preliminary matching mask image that is closest to the motion law data of the target imaging image is determined and used as the first mask image; The preliminary matching mask images other than the first mask image in the preliminary matching mask set are used as the second mask image; wherein, the second position data is determined on the second mask image based on the first position data by template matching.

7. The method according to any one of claims 1 to 5, characterized in that, The step of generating a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image includes: Registration is performed based on the target matching mask image and the target angiography image to obtain a registered mask image; the registered mask image is obtained by subtracting the registered mask image from the target angiography image; or The target angiography image is obtained by subtracting the target matching mask image from the target angiography image.

8. The method according to claim 1, characterized in that, The preliminary matching mask images are multiple images, and these multiple preliminary matching mask images constitute a preliminary matching mask set; the preliminary matching mask set is determined by the following method: Obtain the mask image sequence; Determine the third position data of the key points of the target imaging image and the fourth position data of the key points of each frame of the masked image in the masked image sequence; Based on the third position data and the fourth position data, determine the image similarity data between each frame of the mask image in the mask image sequence and the target imaging image; The preliminary matching mask set is determined in the mask image sequence based on the image similarity data.

9. A device for generating vascular subtraction images, characterized in that, The apparatus for implementing the method of any one of claims 1 to 8 comprises: The image acquisition module is used to acquire the target contrast image and the preliminary matching mask image that matches the motion pattern of the target contrast image; An image fusion module is used to process the preliminary matching mask image to obtain a target matching mask image whose background is similar to that of the target imaging image. The subtraction image generation module is used to generate a vascular subtraction image corresponding to the target angiography image based on the target matching mask image and the target angiography image.

10. A medical imaging device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images