Ultrasonic image segmentation method and device, electronic equipment and storage medium

Abstract

The invention provides an ultrasonic image segmentation method and device, electronic equipment and a storage medium. The method comprises the steps that an ultrasonic image sequence comprising a plurality of ultrasonic images at different moments is acquired, the ultrasonic image sequence comprises the ultrasonic images at different moments, and the ultrasonic images comprise ultrasonic images at a plurality of preset time periods; for at least one target ultrasonic image in the ultrasonic image sequence, determining a target query vector corresponding to a target tissue in the target ultrasonic image according to an initial query vector corresponding to the tissue in the ultrasonic image sequence and a preset time period in which the target ultrasonic image is located, the initial query vector is determined based on features of tissues in the ultrasonic image sequence in a plurality of ultrasonic images in a corresponding preset time period; and performing image segmentation on the target ultrasonic image according to the target query vector to obtain a segmented image of the target tissue in the target ultrasonic image. According to the scheme, the segmentation effect of the target tissue in the target ultrasonic image of the ultrasonic image sequence can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to an ultrasonic image segmentation method, an ultrasonic image segmentation device, an electronic device, a storage medium and a computer program product. Background Art

[0002] Ultrasound examination is a new examination technology in recent years. It uses high-frequency ultrasound to diagnose musculoskeletal system diseases. It can clearly show superficial soft tissue structures such as muscles, tendons, ligaments, and peripheral nerves, as well as their lesions, such as structural abnormalities caused by inflammation, tumors, injuries, and deformities. Combined with relevant medical history and clinical symptoms, most cases can be accurately diagnosed. High-frequency ultrasound's ability to display soft tissue lesions is comparable to that of magnetic resonance imaging, and it can finely distinguish the anatomical structures of muscles and superficial nerves.

[0003] During ultrasound examination, many biological tissues are anisotropic, such as tendons and nerve tissues. Anisotropic artifacts are common artifacts during ultrasound examinations of tendons, nerve tissues, etc. The angle between the incident ultrasound beam and the interface is clearly related to the strength of the echo signal. When the incident beam is perpendicular to the interface, the echo is the strongest. As the incident angle between the incident beam and the interface deviates from the perpendicular state, most of the echo signals will deviate from the incident line and cannot return to the probe, resulting in a weakened echo signal received by the probe.

[0004] When performing ultrasound examination on target tissue, in some application scenarios, it is expected that the target tissue can be accurately segmented in each frame of ultrasound image to determine the morphology of the target tissue. However, anisotropy can cause drastic changes in the target tissue in the ultrasound image, which is not conducive to the image segmentation of the ultrasound image and will affect the segmentation effect of the ultrasound image. If only relying on a general image segmentation network, it is difficult to accurately segment the target tissue in the ultrasound image. Summary of the invention

[0005] The present invention has been made in view of the above-mentioned problems.

[0006] According to a first aspect of the present invention, a method for ultrasonic image segmentation is provided. The method comprises: acquiring an ultrasonic image sequence, wherein the ultrasonic image sequence comprises a plurality of ultrasonic images at different times, and the plurality of ultrasonic images comprises ultrasonic images at a plurality of preset time periods; for at least one target ultrasonic image in the ultrasonic image sequence, determining a target query vector corresponding to a target tissue in the target ultrasonic image according to an initial query vector corresponding to a tissue in the ultrasonic image sequence and a preset time period in which the target ultrasonic image is located, wherein the initial query vector is determined based on features of the tissue in the ultrasonic image sequence in a plurality of ultrasonic images in the corresponding preset time period; and performing image segmentation on the target ultrasonic image according to the target query vector to obtain a segmented image of the target tissue in the target ultrasonic image.

[0007] Exemplarily, the method further includes: determining all tissues included in the ultrasound image sequence; and determining an initial query vector corresponding to each tissue in each preset time period based on features of each tissue in all ultrasound images in each preset time period.

[0008] Exemplarily, the moment corresponding to the target ultrasound image is the first moment, and the preset time period in which the first moment is located is the first preset time period; determining the target query vector corresponding to the target tissue in the target ultrasound image according to the initial query vector corresponding to the tissue in the ultrasound image sequence and the preset time period in which the target ultrasound image is located includes: for each preset time period, determining a moment in the preset time period as a reference moment; determining a target time period among the multiple preset time periods according to the reference moment of the first preset time period and the first moment, wherein the target time period includes the first preset time period and a preset time period adjacent to the first preset time period, and the adjacent preset time period is closer to the first moment than to the reference moment of the first preset time period; calculating the target query vector corresponding to the target tissue in the target ultrasound image according to the initial query vector corresponding to each tissue in the target time period, the reference moment and the first moment.

[0009] Exemplarily, the method of calculating the target query vector corresponding to the target tissue in the target ultrasound image based on the initial query vector corresponding to each tissue in the target time period, the reference moment and the first moment, includes: for any target tissue in the target ultrasound image, determining a first coefficient of the initial query vector corresponding to the target tissue in the first preset time period based on a first time interval between the reference moment in the adjacent preset time period and the first moment; determining a second coefficient of the initial query vector corresponding to the target tissue in the adjacent preset time period based on a second time interval between the reference moment in the first preset time period and the first moment; determining a final query vector as the target query vector corresponding to the target tissue in the target ultrasound image based on the first coefficient, the second coefficient, the initial query vector corresponding to the target tissue in the first preset time period, and the initial query vector corresponding to the target tissue in the adjacent preset time period.

[0010] Exemplarily, any reference moment determined is a middle moment of a preset time period in which the reference moment is located.

[0011] Exemplarily, the method further includes: determining the division time of the ultrasound image sequence according to the examination item category of the ultrasound examination and / or the examination time corresponding to the examination item category, the examination item category corresponding to the tissue in the ultrasound image sequence; accordingly, after acquiring the ultrasound image sequence, it further includes: dividing multiple ultrasound images at different times according to the division time corresponding to the acquired ultrasound image sequence to obtain the ultrasound images of the multiple preset time periods.

[0012] Exemplarily, the division duration of the ultrasound image sequence is determined according to the examination item category of the ultrasound examination and / or the examination duration corresponding to the examination item category, including: if the examination item category is an examination item corresponding to the heart, the division duration of the ultrasound image sequence is determined as the first division duration; if the examination item category is an examination item corresponding to obstetrics and gynecology, an examination item corresponding to the thyroid, an examination item corresponding to the breast, an examination item corresponding to the musculoskeletal shoulder joint, an examination item corresponding to the musculoskeletal ankle joint, or an examination item corresponding to the abdominal organs, the division duration of the ultrasound image sequence is determined as the second division duration; wherein the first division duration is greater than the second division duration.

[0013] Exemplarily, the duration of the preset time period is less than or equal to a set value.

[0014] Exemplarily, the target tissue includes muscle tissue and / or bone.

[0015] According to a second aspect of the present invention, there is also provided an ultrasonic image segmentation device, comprising:

[0016] A receiving module, used for acquiring an ultrasound image sequence, wherein the ultrasound image sequence includes a plurality of ultrasound images at different moments, and the plurality of ultrasound images include ultrasound images at a plurality of preset time periods;

[0017] a determination module, for at least one target ultrasound image in the ultrasound image sequence, determining a target query vector corresponding to a target tissue in the target ultrasound image according to an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on features of a plurality of ultrasound images of the tissue in the ultrasound image sequence in a corresponding preset time period;

[0018] The segmentation module is used to perform image segmentation on the target ultrasound image according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image.

[0019] According to a third aspect of the present invention, there is also provided an electronic device, comprising: a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions are used to execute the above-mentioned ultrasound image segmentation method when the processor is running.

[0020] According to a fourth aspect of the present invention, a storage medium is further provided, on which program instructions are stored, and the program instructions are used to execute the above-mentioned ultrasound image segmentation method when running.

[0021] According to a fifth aspect of the present invention, there is also provided a computer program product, comprising computer program instructions, wherein the computer program instructions are used to execute the above-mentioned ultrasound image segmentation method when running.

[0022] In the above technical solution, an ultrasound image sequence including multiple ultrasound images at different times is obtained, wherein the multiple ultrasound images include ultrasound images of multiple preset time periods, and then for at least one target ultrasound image in the ultrasound image sequence, a target query vector corresponding to the target tissue in the target ultrasound image is determined according to an initial query vector corresponding to the tissue in the ultrasound image sequence and the preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on the features of the tissue in the ultrasound image sequence in multiple ultrasound images in the corresponding preset time period, and then the target ultrasound image is segmented according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image. Because the initial vector is determined based on the features in multiple ultrasound images, the tissues in the ultrasound images within the preset time period share an initial query vector, which will link the time sequence information of multiple ultrasound images. And because the target query vector is determined based on the initial query vector, when the target ultrasound image is segmented, the influence of the anisotropy of the target tissue on the segmentation of the target ultrasound image can be reduced, and the segmentation effect of the target tissue in each target ultrasound image in the ultrasound image sequence can be improved.

[0023] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and other purposes, features and advantages of the present invention will become more apparent by describing the embodiments of the present invention in more detail in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0025] Figure 1 A schematic diagram showing an ultrasound image in an ultrasound image sequence according to an embodiment of the present invention;

[0026] Figure 2 A schematic flow chart of an ultrasound image segmentation method according to an embodiment of the present invention is shown;

[0027] Figure 3 A schematic flow chart of determining an initial query vector according to an embodiment of the present invention is shown;

[0028] Figure 4 A schematic flow chart of determining a target query vector corresponding to a target tissue in a target ultrasound image according to an embodiment of the present invention is shown;

[0029] Figure 5 A schematic diagram showing a method of determining a first target time period from among a plurality of preset time periods according to an embodiment of the present invention is shown;

[0030] Figure 6 A schematic flow chart showing the calculation of a target query vector corresponding to a target tissue in a target ultrasound image according to an initial query vector corresponding to each tissue in a first target time period, a reference time and a first time;

[0031] Figure 7 A schematic flow chart of determining a target query vector corresponding to a target tissue in the target ultrasound image according to yet another embodiment of the present invention is shown;

[0032] Figure 8 A schematic diagram showing a method of determining a second target time period from a plurality of preset time periods according to an embodiment of the present invention is shown;

[0033] Fig. 9 A schematic flow chart of calculating a target query vector of a target ultrasound image according to at least an initial query vector corresponding to each preset time period within a target time period, a first change moment and a first moment according to an embodiment of the present invention is shown;

[0034] Fig.10 A schematic flow chart of calculating a target query vector of a target ultrasound image according to an initial query vector corresponding to each preset time period within a target time period, a second change moment and a first moment according to another embodiment of the present invention is shown;

[0035] Fig.11 A schematic diagram showing an ultrasound image segmentation method according to an embodiment of the present invention is shown;

[0036] Fig.12 A schematic block diagram of an ultrasound image segmentation device according to an embodiment of the present invention is shown;

[0037] Fig.13 A schematic block diagram of an electronic device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical scheme and advantages of the present invention more obvious, the exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described herein. Based on the embodiments of the present invention described in the present invention, all other embodiments obtained by those skilled in the art without creative work should fall within the protection scope of the present invention.

[0039] When segmenting the target object in the image, the image area where the target object is located can usually be determined based on image information such as the gradient information and edge information of the image, and then the segmented image can be obtained according to the image segmentation algorithm. You can also use a general image segmentation model to identify the target object in the image and then output the segmented image. Although both methods can segment the target object in the image, the segmentation effect depends on the clarity of the target object in the image. In ultrasound images, because the target tissue may be anisotropic, it will cause drastic changes in the target tissue in the ultrasound image. Using existing image algorithms and image segmentation models to segment this type of ultrasound image will affect the accuracy of the target tissue in the segmented image, which is not conducive to identifying the target tissue therein.

[0040] Figure 1 A schematic diagram of an ultrasound image in an ultrasound image sequence according to an embodiment of the present invention is shown.

[0041] like Figure 1 As shown in the figure, the three ultrasound images are ultrasound images of the thumb flexor tendon (wrist) at similar times. However, due to anisotropy, the thumb flexor tendon (wrist) is fully displayed in the first ultrasound image, and the same cross section of the thumb flexor tendon (wrist) is displayed more shallowly in the second ultrasound image. In the third ultrasound image, due to anisotropy, the same cross section of the thumb flexor tendon (wrist) is completely invisible and has no echo. At this time, the existing image algorithm and image segmentation model are used to segment this type of ultrasound image, and the thumb flexor tendon (wrist) in the second image and the third image cannot be effectively segmented.

[0042] In order to at least partially solve the above problems, an ultrasonic image segmentation method is proposed. The method obtains an ultrasonic image sequence including multiple ultrasonic images at different times, wherein the multiple ultrasonic images include ultrasonic images of multiple preset time periods, and then for at least one target ultrasonic image in the ultrasonic image sequence, according to the initial query vector corresponding to the tissue in the ultrasonic image sequence and the preset time period in which the target ultrasonic image is located, the target query vector corresponding to the target tissue in the target ultrasonic image is determined, wherein the initial query vector is determined based on the features of the tissue in the ultrasonic image sequence in multiple ultrasonic images in the corresponding preset time period; then the target ultrasonic image is segmented according to the target query vector to obtain a segmented image of the target tissue in the target ultrasonic image. In this way, the target query vector can be linked to the timing information of multiple ultrasonic images, and the segmentation effect of each tissue in the ultrasonic image sequence can be improved. It should be noted that in the embodiment of the present application, the multiple can be at least two.

[0043] Figure 2 FIG. 4 is a schematic flow chart of an ultrasound image segmentation method according to an embodiment of the present invention. Figure 2 As shown, the ultrasound image segmentation method may include steps S110 to S130.

[0044] In step S110 , an ultrasound image sequence is acquired, where the ultrasound image sequence includes a plurality of ultrasound images at different moments, and the plurality of ultrasound images include ultrasound images at a plurality of preset time periods.

[0045] The ultrasonic image sequence may be an ultrasonic video, or an image sequence consisting of ultrasonic images at corresponding moments. The ultrasonic images in the ultrasonic image sequence may be RGB images, grayscale images, or images in other formats. These ultrasonic images may also be images of any suitable size and resolution. These ultrasonic images may be original images directly acquired by an ultrasonic device, or images after a preprocessing operation has been performed on the original images. The preprocessing operation may include operations such as digitization, geometric transformation, normalization, and filtering of the original images to obtain an ultrasonic image sequence for performing subsequent steps of the ultrasonic image segmentation method provided by the present invention.

[0046] Each ultrasonic image in the ultrasonic image sequence corresponds to a moment, which can be the acquisition moment of the ultrasonic image or a specified moment. The multiple moments at which the multiple ultrasonic images are located can be used as a preset time period, so that the moments at which the ultrasonic images in the ultrasonic image sequence are located are divided into multiple time periods. The preset time period can be a preset time period for continuous moments or a preset time period for discontinuous moments. It can be understood that when each preset time period has the same duration and the moments therein are continuous moments, the computational complexity can be reduced. Different preset time periods may not intersect with each other, or there may be overlaps. For example, the preset time period may include 0s-5s and 5s-10s, and the two preset time periods overlap at the time intersection at 5s. When the preset time periods do not intersect with each other, the moment of each ultrasonic image can be made to be within only one preset time period, avoiding conflicts when performing calculations involving preset time periods.

[0047] At least one tissue may be ultrasonically scanned to obtain an ultrasonic video within a period of time, as the ultrasonic image sequence, and each frame in the ultrasonic video may be used as an ultrasonic image to be analyzed and processed. The analysis and processing may include image segmentation, feature extraction, filtering and other analysis and processing operations. Among them, each ultrasonic image in the ultrasonic video may be a section image obtained by scanning. For an ultrasonic scan, at least one tissue adjacent to a certain part or region may be scanned, for example, the wrist joint may be scanned to obtain an ultrasonic section corresponding to the longitudinal section of the wrist joint, the cross section of the first cavity, and the cross sections of the second to fifth cavities to form an ultrasonic video; or at least one tissue in different parts may be scanned separately, for example, the liver, pancreas, and kidney may be scanned in turn to form an ultrasonic video. During the ultrasonic scanning process, the scanning time corresponding to different tissues may be the same or different.

[0048] In step S120, for at least one target ultrasound image in the ultrasound image sequence, a target query vector corresponding to the target tissue in the target ultrasound image is determined based on an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on features of the tissue in the ultrasound image sequence in multiple ultrasound images in the corresponding preset time period.

[0049] The preset time period at which the moment corresponding to the ultrasound image is located is the preset time period in which the ultrasound image is located. The target ultrasound image may be any ultrasound image in the ultrasound image sequence. The embodiment of the present application may perform subsequent image segmentation processing steps on at least one target ultrasound image.

[0050] It should be noted that when executing the processing step of determining the target query vector, only one target ultrasound image may be processed at a time, or multiple target ultrasound images may be processed at the same time. For example, multiple target ultrasound images within the same preset time period may be processed at the same time. When multiple target ultrasound images are processed at one time, the multiple target ultrasound images may be treated as an image set, that is, as a whole, and the corresponding steps are performed with one image set as a processing object. Therefore, the same target tissue in the target ultrasound images in one image set corresponds to the same target query vector.

[0051] It is understandable that multiple types of tissues such as tendons, bones, and nerve tissues may exist simultaneously in the ultrasound image. According to the ultrasound scanning range corresponding to the ultrasound image, the tissues in the ultrasound image of the ultrasound image sequence can be determined. For example, when the ultrasound scanning range corresponding to the ultrasound image is only the liver, the tissues in the ultrasound image may include the liver, gallbladder, etc., but there will be no limb bones. When the ultrasound scanning range corresponding to the ultrasound image is only the limb joints, the tissues in the ultrasound image may include tendons, bones, etc., but there will be no liver, heart, etc. The features of the tissue in each image of the ultrasound image sequence can be determined by a feature analysis algorithm. Based on the features of the tissue in each ultrasound image and the preset time period in which each ultrasound image is located, the features of the tissue in the ultrasound image of each preset time period can be determined. The target tissue is at least a part of the tissue in the ultrasound image sequence. The corresponding preset time period corresponds to the tissue in the ultrasound image sequence. For each tissue in the ultrasound image sequence, the tissue should exist in at least one ultrasound image of the corresponding preset time period corresponding to the tissue.

[0052] Exemplarily, the duration of the preset time period is less than or equal to the set value. The set value can be 10 seconds or other values. According to the characteristics of ultrasound, doctors will only stay for less than 5 seconds for organs in the same position most of the time, and then scan other parts. Unless a lesion is seen, detailed observation is required, which can be up to 10 seconds. Therefore, the duration of the preset time period should not be set too long to avoid the existence of only one ultrasound image within a preset time period in the ultrasound image sequence, so that only one preset time period can be determined, which is not conducive to image segmentation. For example, for an ultrasound video of 1800 frames with a frame rate of 30 frames, its duration is 60 seconds, and the corresponding moment can be divided into 6 preset time periods, that is, every 10 seconds as a preset time period.

[0053] For example, in each preset time period, feature analysis can be performed on all tissues included in the ultrasound image sequence to determine the features of each tissue in the ultrasound image in each preset time period. Then, based on the features of each tissue in the ultrasound image in each preset time period, the initial query vector corresponding to each tissue in each preset time period is determined. For example, there are 10 tissues included in the ultrasound image sequence, and there are 6 preset time periods. In each preset time period, the corresponding initial query vector is obtained based on the features of these 10 tissues, so that 60 initial query vectors can be obtained. Further, if a certain tissue A does not exist in a preset time period T1, the initial query vector of tissue A in the preset time period T1 can be 0 or a null value. Then, according to the preset time period in which the target ultrasound image is located and the initial query vector corresponding to each tissue in each preset time period, the target query vector corresponding to the target tissue in the target ultrasound image is determined.

[0054] For example, in each preset time period, feature analysis can be performed only on several tissues present in the ultrasound image of the preset time period to determine the features of each tissue in the ultrasound image of the preset time period; at this time, the tissue should exist in at least one ultrasound image of the preset time period; then, the initial query vector of each tissue in the preset time period can be determined based on the features of the tissue in the corresponding ultrasound image. For the target ultrasound image in the preset time period, based on the initial query vector of the target tissue in the preset time period, the target query vector corresponding to the target tissue in the target ultrasound image can be determined. For example, for an ultrasound image sequence with tendons and bones, by performing feature analysis on the tendons and bones in the ultrasound image sequence, ultrasound images from which the features of the tendons and bones can be extracted can be determined, and then the preset time periods in which these ultrasound images are located can be determined as the corresponding time periods corresponding to the tendons and bones. At least one initial query vector can be determined based on the features of the tendons in multiple ultrasound images in the corresponding time period, and at least one initial query vector can be determined based on the features of the bones in multiple ultrasound images in the corresponding time period. If the preset time period of the target ultrasound image is the preset time period of the tendon or bone, the initial query vectors corresponding to the bones and tendons in the preset time period of the target ultrasound image can be determined as the target query vectors corresponding to the tendons and bones in the target ultrasound image. Further, for the ultrasound image sequence including preset time period 1, preset time period 2, preset time period 3, and preset time period 4, the ultrasound image sequence includes three tissues: tendons, bones, and cysts. Tendons and bones can extract features in both preset time period 1 and preset time period 2, and cysts cannot extract features in preset time period 1 and preset time period 2, which means that tendons and bones exist in the ultrasound images in preset time period 1 and preset time period 2, but no cysts exist, and preset time period 1 and preset time period 2 are used as the corresponding time periods corresponding to tendons and bones. Then, according to the features of the tendons and bones in the ultrasound image of preset time period 1, the initial query vectors corresponding to the tendons and bones in preset time period 1 can be determined. Similarly, according to the features of the tendons and bones in the ultrasound image of preset time period 2, the initial query vectors corresponding to the tendons and bones in preset time period 2 can be determined. That is, when determining the initial query vectors in preset time periods 1 and 2, only the initial query vectors corresponding to the tendons and bones are determined, and there is no need to determine the initial query vector corresponding to the cyst. For the target ultrasound image in preset time period 1, the target query vectors corresponding to the tendons and bones in the target ultrasound image can be determined based on the initial query vectors corresponding to the tendons and bones in preset time period 1 and the initial query vectors corresponding to the tendons and bones in preset time period 2, respectively, without determining the target query vector corresponding to the cyst. Since the tissues in the ultrasound images in preset time periods 3 and 4 do not include tendons and bones, the features of the tendons and bones cannot be extracted, and they will not participate in the subsequent calculation of the target query vectors corresponding to the tendons and bones.In addition, even if there are tissues other than tendons and bones (such as the aforementioned cysts) in the ultrasound images of preset time periods 1, 2, 3, and 4, the features of the tissues other than tendons and bones may not be extracted. In this way, while ensuring that the target query vector corresponding to the target tissue in the target ultrasound image can be determined, it is not necessary to perform feature extraction on all tissues in the ultrasound image sequence.

[0055] Optionally, when the corresponding preset time period includes multiple preset time periods including the preset time period in which the target ultrasound image is located, the determined initial query vectors corresponding to the tendon and the bone are also multiple. At this time, the target query vector corresponding to the tendon and the bone in the target ultrasound image can also be determined based on the initial query vectors corresponding to the tendon and the bone in the multiple preset time periods.

[0056] For example, the initial query vector corresponding to the target tissue in the preset time period of the target ultrasound image can be directly used as the target query vector. According to the characteristics of ultrasound, doctors usually stay for a period of time to scan organs in the same position, so the ultrasound scanning range corresponding to the ultrasound images collected within a preset time period will not change too much. Although the target tissues in these ultrasound images may appear to be lighter or disappear due to anisotropy, the morphology and relative position between different target tissues do not change much. Because the target ultrasound images in the same preset time period share the same query vector, the segmentation effect of the target tissue in the preset time period is basically similar, so the initial query vector corresponding to the target tissue in the preset time period of the target ultrasound image can be directly used as the target query vector.

[0057] For example, the target tissue in the ultrasound image in the same preset time period may also undergo drastic changes due to the large change in the ultrasound scanning range. For example, in the preset time period 0s-5s, the ultrasound scanning range in 0s-3s is the heart, and the ultrasound scanning range in 3s-5s is the liver, then the target tissue in the ultrasound image in this part of the time period is obviously different. If the target ultrasound image in the preset time period is still segmented according to the initial query vector, the segmentation effect of the target tissue in the target ultrasound image will be affected. For example, the target query vector corresponding to the target tissue in the target ultrasound image can be determined in combination with the change of the target tissue in the ultrasound image sequence and the initial query vector corresponding to the preset time period in which the target ultrasound image is located. For example, the target query vector corresponding to the target tissue in the target ultrasound image can be determined according to the change moment of the target tissue in the ultrasound image sequence, the moment corresponding to the target ultrasound image, and the initial query vector corresponding to the preset time period in which the target tissue is located at the change moment. In this way, the change of the target tissue in the ultrasound image in the same time period can be taken into account, and it is ensured that the determined target query vector can provide a better image segmentation effect.

[0058] Exemplarily, the initial query vectors corresponding to the same target organization in different preset time periods are different.

[0059] In step S130, the target ultrasound image is segmented according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image.

[0060] According to the target query vector corresponding to the target ultrasound image, target detection can be performed on the target tissue in the target ultrasound image, and after the target tissue corresponding to the query vector is detected, the target tissue is segmented, and the image area of ​​the non-target tissue is used as the background in the segmented image. For example, when all ultrasound images of the ultrasound image sequence are processed, the target tissue in each ultrasound image can be segmented more accurately using the above steps.

[0061] For example, the target query vector corresponding to the target tissue in the target ultrasound image and the target ultrasound image can be input into an image segmentation model to output a segmented image of the target ultrasound image. The image segmentation model can be an image segmentation model for instance segmentation, such as Mask R-CNN, YOLOV8, RTMDET, DeepLab, etc. In other embodiments, the target ultrasound image can also be segmented by a certain image segmentation unit in the image segmentation model.

[0062] Exemplarily, the target tissue includes muscle tissue and / or bone. According to different types of bone and / or muscle tissue, target query vectors corresponding to all types of bone and / or muscle tissue can be determined. After the target ultrasound image containing bone and / or muscle tissue is input into the image segmentation model, the image segmentation model can detect the type of bone and / or muscle tissue, and segment the bone and / or muscle tissue in the target ultrasound image according to the target query vector to obtain a segmented image of the target ultrasound image.

[0063] During ultrasound examination, the anisotropy of muscle tissue and bone is usually strong, and the changes between ultrasound images are more intense. The above steps can be used to connect the time sequence information of multiple ultrasound images to more accurately segment the muscle tissue and / or bone in the ultrasound image.

[0064] In the above technical solution, an ultrasound image sequence is obtained, the ultrasound image sequence includes multiple ultrasound images at different times, and the multiple ultrasound images include ultrasound images of multiple preset time periods. Then, for at least one target ultrasound image in the ultrasound image sequence, a target query vector corresponding to the target tissue in the target ultrasound image is determined according to the initial query vector corresponding to the tissue in the ultrasound image sequence and the preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on the features of the tissue in the ultrasound image sequence in multiple ultrasound images in the corresponding preset time period, and then the target ultrasound image is segmented according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image. Because the initial vector is determined based on the features in multiple ultrasound images, the tissues in the ultrasound images within the preset time period share an initial query vector, which will link the time sequence information and associated features of multiple ultrasound images. And because the target query vector is determined based on the initial query vector, when the target ultrasound image is segmented, the influence of the anisotropy of the target tissue on the segmentation of the target ultrasound image can be reduced, and the segmentation effect of the target tissue in each target ultrasound image in the ultrasound image sequence can be improved.

[0065] Figure 3 FIG. 4 shows a schematic flow chart of determining an initial query vector according to an embodiment of the present invention. Figure 3 As shown, the above step S120 may include steps S310 to S320.

[0066] In step S310 , all tissues included in the ultrasound image sequence are determined.

[0067] During ultrasound scanning, the scanned object often changes. For example, when there are 10 tissues in the ultrasound image sequence, the tissues in each preset time period will only be part of the 10 tissues because of the change in the ultrasound scanning range. In ultrasound images of different preset time periods, the morphology of the same tissue may change, such as becoming larger or smaller, and the same tissue may also appear and disappear. Although the morphology of the same tissue in different ultrasound images may change, the features are usually similar, so it can be regarded as the same tissue.

[0068] For example, the features of all tissues contained in the ultrasound image sequence may be extracted according to a feature extraction algorithm, and then the features of all tissues may be compared with standard features corresponding to all tissues, and all tissues may be determined according to the similarity between the features.

[0069] For example, an image segmentation algorithm may be used to segment all tissues contained in an ultrasound image sequence, and then the morphology of all segmented tissues may be compared with the standard morphology of all tissues, and all tissues may be determined based on similarity.

[0070] For example, a neural network model may be used to automatically identify all tissues contained in an ultrasound image sequence based on the characteristics of each tissue.

[0071] The purpose of determining all tissues is to determine the number of initial query vectors. Therefore, in the aforementioned implementation method of determining all tissues, it is not necessary to clearly know the category of each tissue and the regional boundaries and duration of the tissue. It is only necessary to determine the number of tissues present in the ultrasound image sequence, that is, if there is a feature corresponding to a certain tissue, the number of tissues is increased by one.

[0072] In step S320, based on the features of each tissue in all ultrasound images in each preset time period, an initial query vector corresponding to each tissue in each preset time period is determined.

[0073] In ultrasound images of different preset time periods, the same tissue may change, so the initial query vector corresponding to the same tissue may be different. In ultrasound images within a single preset time period, the same tissue will not change too much, so the initial query vector corresponding to the same tissue in all ultrasound images of a single preset time period can be the same, that is, the same tissue in all ultrasound images of a single preset time period shares the same initial query vector. When determining the initial query vector corresponding to each tissue, although the morphology of the same tissue in different ultrasound images may change, they are all regarded as the same tissue.

[0074] For example, for all ultrasound images in any preset time period, a feature extraction algorithm can be used to extract the features of each tissue from each ultrasound image in the preset time period. It can be understood that for any tissue, when the tissue appears in multiple ultrasound images in the preset time period, multiple features of the tissue can be extracted, and then the initial query vector corresponding to the tissue in the preset time period can be determined based on the multiple extracted features of the tissue. The initial query vector determined in this way can be used to query the tissue that has appeared in all ultrasound images in the preset time period. Because the morphology of the same tissue in ultrasound images of different preset time periods may be too different, and the time information of each preset time period is also different, the initial query vectors corresponding to each tissue in different preset time periods are different.

[0075] In the above technical solution, all tissues included in the ultrasound image sequence are determined, and then the initial query vector corresponding to each tissue in each preset time period is determined based on the features of each tissue in all ultrasound images in each preset time period. Determining the features of each tissue first, and then determining each initial query vector, can simply and accurately determine the initial query vector corresponding to each tissue in each preset time period, and is also conducive to further determining the target query vector corresponding to the target tissue in the target ultrasound image. And because the initial query vector is determined based on the features in multiple ultrasound images, the same tissue in the ultrasound images within the preset time period shares an initial query vector, and the timing information and associated features of the tissue in multiple ultrasound images can be linked. And because the target query vector is determined based on the initial query vector, when the target query vector is determined using the initial query vector determined in the above steps, the influence of the anisotropy of the target tissue on the segmentation of the target ultrasound image can be reduced, and the target query vector determined based on the initial query vector can improve the segmentation effect of the target tissue in each target ultrasound image in the ultrasound image sequence.

[0076] In the above technical solution, all tissues contained in the ultrasound image sequence are taken as research objects. For example, if the ultrasound image sequence contains 10 tissues, the initial query vectors of these 10 tissues will be calculated separately in each preset time period. In some examples, the tissues can also be screened and only part of the tissues are processed. For example, based on the user's selection operation, 8 of the tissues are taken as research objects. Then, the initial query vectors of these 8 tissues can be calculated separately in each preset time period. When processing each target ultrasound image, the target query vector corresponding to these 8 tissues is determined, and then these 8 tissues are segmented from the ultrasound image sequence. In this way, the segmentation of the tissue that the user is concerned about can be achieved, and the processing objects of determining the query vector and the image segmentation process are reduced, which can reduce a certain amount of calculation.

[0077] Exemplarily, the moment corresponding to the target ultrasound image may be the first moment, and the preset time period in which the first moment is located may be the first preset time period. Figure 4 FIG. 4 is a schematic flow chart of determining a target query vector corresponding to a target tissue in a target ultrasound image according to an embodiment of the present invention. Figure 4 As shown, the above step S120 may include steps S410 to S430.

[0078] In step S410, for each preset time period, a moment in the preset time period is determined as a reference moment.

[0079] For example, the reference time may be any time in a preset time period.

[0080] For example, the reference time in each preset time period has the same position in the preset time period. For example, any reference time determined may be the start time, end time, etc. of the preset time period in which the reference time is located.

[0081] Exemplarily, any reference moment determined is the middle moment of the preset time period in which the reference moment is located. For an ultrasound image at the middle moment of the preset time period, the state of the tissue can best represent the state of the tissue in ultrasound images at all moments in the preset time period.

[0082] In step S420, a first target time period is determined among multiple preset time periods based on a reference time of the first preset time period and the first time, wherein the first target time period includes the first preset time period and a preset time period adjacent to the first preset time period, and the adjacent preset time period is closer to the first time than to the reference time of the first preset time period.

[0083] Figure 5 A schematic diagram showing determining a first target time period from a plurality of preset time periods according to an embodiment of the present invention is shown.

[0084] like Figure 5 As shown, the first moment and the reference moment of the first preset period are both within the first preset period, and the reference moment within the same preset period is before the first moment, so the first preset period and the adjacent preset period 1 thereafter can be determined as the first target period. Similarly, when the reference moment within the same preset period is after the first moment, the first preset period and the adjacent preset period thereafter can also be determined as the first target period.

[0085] In step S430, a target query vector corresponding to the target tissue in the target ultrasound image is calculated according to the initial query vector, the reference time and the first time corresponding to each tissue in the first target time period.

[0086] The time interval between the reference moment and the first moment of the first preset time period may be determined first, and then the time interval between the reference moment and the first moment in the adjacent preset time period may be determined. According to the ratio between the two time intervals, the influence of the initial query vector corresponding to the target tissue in the first preset time period and the initial query vector corresponding to the target tissue in the adjacent preset time period on the target query vector may be determined. When calculating the target query vector of the target tissue in the target ultrasound image, the influence may be taken into account to more accurately determine the target query vector of the target tissue.

[0087] In the above technical solution, for each preset time period, a moment in the preset time period is determined as a reference moment, and then a first target time period is determined in multiple preset time periods according to the reference moment and the first moment of the first preset time period, wherein the first target time period includes the first preset time period and a preset time period adjacent to the first preset time period, and the adjacent preset time period is closer to the first moment than to the reference moment of the first preset time period, and then the target query vector corresponding to the target tissue in the target ultrasound image is calculated according to the initial query vector, the reference moment and the first moment corresponding to each tissue in the first target time period. In this way, the influence of tissues in ultrasound images within multiple preset time periods on the segmented image can be taken into account, and a target query vector of the target tissue that is more suitable for segmenting the target ultrasound image can be determined, and the target query vector can improve the segmentation effect of the target tissue in the target ultrasound image.

[0088] Figure 6 FIG. 1 is a schematic flow chart of calculating a target query vector corresponding to a target tissue in a target ultrasound image according to an initial query vector corresponding to each tissue in a first target time period, a reference time and a first time according to an embodiment of the present invention. Figure 6 As shown, the above step S430 may include steps S610 to S630.

[0089] In step S610, for any target tissue in the target ultrasound image, a first coefficient of an initial query vector corresponding to the target tissue in a first preset time period is determined according to a first time interval between a reference moment and a first moment in adjacent preset time periods.

[0090] In step S620, a second coefficient of the initial query vector corresponding to the target organization in an adjacent preset time period is determined according to a second time interval between a reference moment in the first preset time period and the first moment.

[0091] When the reference moment in the adjacent preset time period and the reference moment in the first preset time period are fixed, the closer the reference moment in the first preset time period is to the first moment, the farther the reference moment in the adjacent preset time period is from the first moment, and the first coefficient will be larger while the second coefficient will be smaller; the farther the reference moment in the first preset time period is to the first moment, the closer the reference moment in the adjacent preset time period is to the first moment, and the first coefficient will be smaller while the second coefficient will be larger.

[0092] For example, the total duration of the first time interval and the second time interval may be determined first, and then the proportion of the first time interval in the total duration may be determined as the first coefficient, and the proportion of the second time interval in the total duration may be determined as the second coefficient.

[0093] The first coefficient represents the influence of the initial query vector corresponding to the target tissue in the first preset time period on the target query vector corresponding to the target tissue in the target ultrasound image. The second coefficient represents the influence of the initial query vector corresponding to the target tissue in the adjacent preset time period on the target query vector corresponding to the target tissue in the target ultrasound image.

[0094] In step S630, a final query vector is determined as a target query vector corresponding to the target tissue in the target ultrasound image based on the first coefficient, the second coefficient, the initial query vector corresponding to the target tissue in the first preset time period, and the initial query vector corresponding to the target tissue in the adjacent preset time period.

[0095] The total duration of the first time interval and the second time interval may be determined first.

[0096] When the first coefficient is the proportion of the first time interval in the total time length, and the second coefficient is the proportion of the second time interval in the total time length, the product of the first coefficient and the initial query vector corresponding to the target tissue in the first preset time period can be used as the third query vector, and the product of the second coefficient and the initial query vector corresponding to the target tissue in the adjacent preset time period can be used as the fourth query vector. Then, the third query vector and the fourth query vector are fitted to obtain a final query vector as the target query vector corresponding to the target tissue in the target ultrasound image.

[0097] The above-mentioned fitting method may be addition, multiplication, concatenation, etc.

[0098] For example, the target query vector Q corresponding to the target tissue in the target ultrasound image may be determined according to the following formula 1:

[0099] Q = Q0*|T2-t| / T + Q1*|t-T1| / T Formula 1

[0100] Among them, Q0 represents the initial query vector corresponding to the target organization in the first preset time period, Q1 represents the initial query vector corresponding to the target organization in the adjacent preset time period, T1 represents the reference time of the first preset time period, T2 represents the reference time of the adjacent preset time period, t represents the first time, and T represents the total duration of the first time interval and the second time interval.

[0101] It is understandable that there may be multiple target tissues in the target ultrasound image. The above steps are performed for each target tissue to determine the target query vector corresponding to each target tissue in the target ultrasound image, thereby determining the target query vector corresponding to the target tissue in the target ultrasound image.

[0102] In the above technical solution, for any target tissue in the target ultrasound image, according to the first time interval between the reference moment and the first moment in the adjacent preset time period, the first coefficient of the initial query vector corresponding to the target tissue in the first preset time period is determined, and then according to the second time interval between the reference moment and the first moment in the first preset time period, the second coefficient of the initial query vector corresponding to the target tissue in the adjacent preset time period is determined, and then according to the first coefficient, the second coefficient, the initial query vector corresponding to the target tissue in the first preset time period, and the initial query vector corresponding to the target tissue in the adjacent preset time period, the final query vector is determined as the target query vector corresponding to the target tissue in the target ultrasound image. In this way, the influence of the initial query vectors corresponding to any target tissue in multiple preset time periods on the target query vector corresponding to the target tissue in the target ultrasound image can be accurately determined according to the relationship between the reference moment and the first moment. For the target tissue spanning more than one preset time period, the information loss caused by the truncation of the preset time period in the initial query vector can be avoided as much as possible, and the target tissue in the target ultrasound image can be more accurately segmented based on the target query vector determined here.

[0103] Exemplarily, the moment corresponding to the target ultrasound image is the first moment. Figure 7 FIG. 4 is a schematic flow chart of determining a target query vector corresponding to a target tissue in a target ultrasound image according to another embodiment of the present invention. Figure 7 As shown, the above step S120 may include steps S710 to S730.

[0104] In step S710, for any target tissue in the target ultrasound image, the appearance time and disappearance time of the target tissue in the ultrasound image sequence are determined.

[0105] The appearance and disappearance moments of the target tissue in the ultrasound image sequence can be regarded as the change moments of the target tissue. For example, for an ultrasound video with 20 frames per second, there is a bone in the ultrasound image frame at the 3rd second, but the bone disappears in the ultrasound video frame at the next moment, the 3.05th second, then the 3.05th second can be recorded as the hour moment of the bone. Similarly, the bone does not exist in the ultrasound image frame at the 2nd second, but appears in the ultrasound video frame at the next moment, the 2.05th second, then the 2.05th second can be recorded as the appearance moment of the bone. For example, the target detection algorithm can be used to determine the change moment of each target tissue in the ultrasound image sequence, thereby determining the change moment of the target tissue.

[0106] Hereinafter, the appearance period is used to represent the preset period in which the appearance moment in the first change moment is located, and the disappearance period is used to represent the preset period in which the disappearance moment in the first change moment is located, for explanation.

[0107] In step S720, a second target time period is determined from multiple preset time periods based on the first moment and the first change moment of the target tissue, wherein the second target time period includes the preset time period in which the first change moment of the target tissue is located, and the first change moment includes the appearance moment of the target tissue that is before the first moment and closest to the first moment and the disappearance moment that is after the first moment and closest to the first moment.

[0108] It can be understood that the closer the first change moment of the target tissue is to the first moment, the higher the correlation between the ultrasound image in the preset time period where the first change moment is located and the target ultrasound image, so the initial query vector corresponding to the target tissue in the preset time period where the first change moment is located has a greater correlation with the target query vector corresponding to the target tissue in the target ultrasound image. The time interval between the appearance moment of each target tissue in one or more adjacent preset time periods before the target ultrasound image and the first moment can be compared, and then the appearance moment with the smallest time interval is selected as the appearance moment in the first change moment. Similarly, the time interval between the disappearance moment of each target tissue in one or more adjacent preset time periods after the target ultrasound image and the first moment can be compared, and then the disappearance moment with the smallest time interval is selected as the disappearance moment in the first change moment.

[0109] Figure 8 A schematic diagram showing determining a second target time period from a plurality of preset time periods according to an embodiment of the present invention is shown.

[0110] like Figure 8 As shown, the first moment is within the first preset time period. The target tissue in the target ultrasound image includes a first target tissue and a second target tissue. For the first target tissue, the target tissue appears at a moment in preset time period 1 and disappears at a moment in preset time period 2. At this time, preset time period 1 and preset time period 2 can both be used as the second target time period for the first target tissue. For the second target tissue, the target tissue appears at a moment in preset time period 1 and disappears at a moment in preset time period 3. At this time, preset time period 1 and preset time period 3 can both be used as the second target time period for the second target tissue. For different target tissues in the target ultrasound image, because the preset time periods in which the appearance moment and the discovery moment are located may be different, the determined second target time period may also be different.

[0111] In step S730, a target query vector corresponding to the target tissue in the target ultrasound image is calculated according to the initial query vector corresponding to the target tissue in the second target time period, the first change time instant and the first time instant.

[0112] The target tissue in the ultrasound image before and after the first change moment has a large change, so the initial query vector corresponding to the first preset time period cannot be directly used as the target query vector at the first moment.

[0113] The time interval between the disappearance moment and the first moment in the first change moment may be determined first, and then the time interval between the appearance moment and the first moment in the first change moment may be determined. According to the ratio between the two time intervals, the influence of the initial query vector corresponding to the target tissue in the appearance period and the initial query vector corresponding to the target tissue in the disappearance period on the target query vector corresponding to the target tissue in the first moment may be determined. When calculating the target query vector corresponding to the target tissue in the target ultrasound image, the influence may be taken into account to more accurately determine the target query vector corresponding to the target tissue.

[0114] In the above technical solution, for any target tissue in the target ultrasound image, the appearance moment and disappearance moment of the target tissue in the ultrasound image sequence are determined; then, according to the first moment and the first change moment of the target tissue, a second target time period is determined in multiple preset time periods, wherein the second target time period includes the preset time period in which the first change moment of the target tissue is located, and the first change moment includes the appearance moment of the target tissue that is located before the first moment and closest to the first moment and the disappearance moment that is located after the first moment and closest to the first moment; then, according to the initial query vector corresponding to the target tissue in the second target time period, the first change moment and the first moment, the target query vector corresponding to the target tissue in the target ultrasound image is calculated. In this way, the influence of the change of the target tissue in the ultrasound image sequence on the image segmentation can be taken into account, and a target query vector that is more suitable for the target ultrasound image can be determined, and the target query vector can improve the segmentation effect of the target tissue in the target ultrasound image.

[0115] Fig. 9 A schematic flow chart of calculating the target query vector of the target ultrasound image according to at least the initial query vector corresponding to each preset time period within the target time period, the first change time and the first time according to an embodiment of the present invention is shown. Fig. 9 As shown, the above step S730 may include steps S910 to S930.

[0116] In step S910, the third coefficient of the initial query vector corresponding to the target organization in the second preset time period is determined according to the third time interval between the appearance moment in the first change moment and the first moment, wherein the second preset time period is the preset time period in which the disappearance moment in the first change moment is located.

[0117] The process of executing step S910 is similar to that of the above step S610. The initial query vector corresponding to the target organization in the second preset time period can be processed similarly as the initial query vector corresponding to the target organization in the first preset time period in the above step S610 to obtain the fourth coefficient, which will not be described in detail here.

[0118] In step S920, the fourth coefficient of the initial query vector corresponding to the target organization in the third preset time period is determined according to the fourth time interval between the disappearance moment in the first change moment and the first moment, wherein the third preset time period is the preset time period in which the appearance moment in the first change moment is located.

[0119] The process of executing step S920 is similar to that of the above step S620. The initial query vector corresponding to the target organization in the third preset time period can be processed similarly as the initial query vector corresponding to the target organization in the adjacent preset time period in the above step S620 to obtain the fourth coefficient, which will not be described in detail here.

[0120] In step S930, a final query vector is determined as a target query vector corresponding to the target tissue in the target ultrasound image based on the third coefficient, the fourth coefficient, the initial query vector corresponding to the target tissue in the second preset time period, and the initial query vector corresponding to the target tissue in the third preset time period.

[0121] The process of executing step S920 is similar to the above step S630 and will not be described in detail here.

[0122] In the above technical solution, according to the third time interval between the appearance moment in the first change moment and the first moment, the third coefficient of the initial query vector corresponding to the target tissue in the second preset time period is determined, wherein the second preset time period is the preset time period where the disappearance moment in the first change moment is located, and then according to the fourth time interval between the disappearance moment in the first change moment and the first moment, the fourth coefficient of the initial query vector corresponding to the target tissue in the third preset time period is determined, wherein the third preset time period is the preset time period where the appearance moment in the first change moment is located, and then according to the third coefficient, the fourth coefficient, the initial query vector corresponding to the target tissue in the second preset time period and the initial query vector corresponding to the target tissue in the third preset time period, the final query vector is determined as the target query vector corresponding to the target tissue in the target ultrasound image. In this way, the influence of the appearance and disappearance of the target tissue in the ultrasound image sequence on the target query vector corresponding to the target tissue in the target ultrasound image can be accurately determined according to the relationship between the first change moment and the first moment, and the target tissue in the target ultrasound image can be more accurately segmented based on the target query vector determined here.

[0123] Exemplarily, the moment corresponding to the target ultrasound image is the first moment, and the preset time period in which the first moment is located is the first preset time period. Fig.10 FIG. 1 is a schematic flow chart showing a method of calculating a target query vector of a target ultrasound image according to an initial query vector corresponding to each preset time period within a target time period, a second change time instant and a first time instant according to another embodiment of the present invention. Fig.10 As shown, it may include steps S1010 to S1030.

[0124] In step S1010, for any target tissue in the target ultrasound image, the fifth coefficient of the initial query vector corresponding to the first preset time period is determined according to the fifth time interval between the second change moment and the first moment of the target tissue in the fourth preset time period, wherein the second change moment is closer to the first moment than the reference moment in the first preset time period, and the second change moment of the target tissue includes the appearance moment or disappearance moment of the target tissue closest to the first moment.

[0125] The process of executing step S1010 is similar to the above-mentioned step S610 and will not be described in detail here.

[0126] In step S1020, a sixth coefficient of the initial query vector corresponding to the target tissue in a fourth preset time period is determined according to a sixth time interval between a reference moment in the first preset time period and the first moment.

[0127] The process of executing step S1020 is similar to the above-mentioned step S620 and will not be described in detail here.

[0128] In step S1030, a final query vector is determined as a target query vector corresponding to the target tissue in the target ultrasound image based on the fifth coefficient, the sixth coefficient, the initial query vector corresponding to the target tissue in the first preset time period, and the initial query vector corresponding to the target tissue in the fourth preset time period.

[0129] The process of executing step S1030 is similar to the above-mentioned step S630 and will not be described in detail here.

[0130] It can be understood that when the appearance time and disappearance time of the target tissue are both before or after the first time, the first change time is incomplete. In this case, the appearance time or disappearance time of the target tissue closest to the first time can be used as the second change time, and combined with the reference time within the first preset time period, the target query vector corresponding to the target tissue in the target ultrasound image is calculated.

[0131] In the above technical solution, for any target tissue in the target ultrasound image, according to the fifth time interval between the second change moment and the first moment of the target tissue in the fourth preset time period, the fifth coefficient of the initial query vector corresponding to the first preset time period is determined, wherein the second change moment in the fourth preset time period is closer to the first moment than the reference moment in the first preset time period, and the second change moment of the target tissue includes the appearance moment or disappearance moment of the target tissue closest to the first moment, and then according to the sixth time interval between the reference moment in the first preset time period and the first moment, the sixth coefficient of the initial query vector corresponding to the fourth preset time period is determined, and then according to the fifth coefficient, the sixth coefficient, the initial query vector corresponding to the target tissue in the first preset time period, and the initial query vector corresponding to the target tissue in the fourth preset time period, the final query vector is determined as the target query vector corresponding to the target tissue in the target ultrasound image. In this way, it can be avoided that the target query vector corresponding to the target tissue in the target ultrasound image cannot be determined due to incomplete distance from the first change moment.

[0132] Ultrasound examination items can be divided into general ultrasound examination items and ultrasound physical examination items; the examination precision requirements of ultrasound physical examination items are often lower than those of general ultrasound examination items. Therefore, the examination time for each scanned object in general ultrasound examination items is often longer. Regardless of whether it is a general ultrasound examination item or an ultrasound physical examination item, it can be subdivided into different examination item categories, such as abdominal ultrasound, cardiac ultrasound, musculoskeletal ultrasound, etc. Based on this, the target division time of the preset time period can be determined in combination with the examination item category and / or the examination time corresponding to the examination item category.

[0133] Exemplarily, for a general ultrasound examination item, the division duration of the ultrasound image sequence is determined according to the ultrasound examination item category and / or the examination duration corresponding to the examination item category, including:

[0134] If the examination item category is an examination item corresponding to the heart, the division time length of the ultrasound image sequence is determined as the first division time length;

[0135] If the examination item category is an examination item corresponding to obstetrics and gynecology, an examination item corresponding to the thyroid gland, an examination item corresponding to the breast, an examination item corresponding to the musculoskeletal shoulder joint, an examination item corresponding to the musculoskeletal ankle joint, or an examination item corresponding to the abdominal organ, the division time length of the ultrasound image sequence is determined to be the second division time length;

[0136] Among them, the duration of the first division is greater than the duration of the second division.

[0137] For example, for a 1-minute ultrasound video, the ultrasound video corresponding to the heart can be divided into 3 preset time periods, and the ultrasound videos corresponding to obstetrics and gynecology, thyroid, breast, musculoskeletal shoulder joint, musculoskeletal ankle joint, abdominal organs, etc. can be divided into 6 segments.

[0138] Exemplarily, the corresponding division time can also be determined based on the historical examination time of the examination items corresponding to obstetrics and gynecology, the examination items corresponding to the thyroid gland, the examination items corresponding to the breast, the examination items corresponding to the musculoskeletal shoulder joints, the examination items corresponding to the musculoskeletal ankle joints, or the examination items corresponding to the abdominal organs, so as to achieve more targeted time division.

[0139] In the above embodiment, different division time lengths are determined for different inspection items. For inspection items that require more observation time or more sophisticated image segmentation processing, a longer division time length is determined. This approach can take into account both the accuracy and processing efficiency of image segmentation.

[0140] For physical examination items, the duration can be shortened based on the division duration corresponding to the aforementioned general ultrasound examination items to improve the efficiency of image segmentation and ultrasound physical examination. The duration shortening process can be to subtract the set difference from the corresponding division duration, or to reduce the corresponding division duration by a certain multiple.

[0141] Exemplarily, the above method further includes step S140: determining a preset time period according to the type of department used to examine the target tissue.

[0142] Different departments need to examine different tissues. For example, orthopedics usually examines bones and muscle tissues around bones, neurology usually examines nerve tissues, and obstetrics and gynecology usually examines fetal information. And different departments have different degrees of precision when performing ultrasound examinations because they need to examine different tissues. When a tissue needs to be examined with higher precision, the probe used to scan the tissue usually needs to remain at the target location for a longer time, and the tissue will not change very frequently. Therefore, fewer preset time periods can be determined based on the time of the ultrasound video sequence, and the duration of each preset time period will also be longer. When a tissue needs to be examined with lower precision, the probe used to scan the tissue usually needs to remain at the target location for a shorter time, and the tissue will not change very frequently. At this time, more preset time periods can be determined based on the time of the ultrasound video sequence, and the duration of each preset time period will also be shorter.

[0143] According to the type of department used to examine the tissue, the ultrasound examination habits of the medical staff in the department can be determined, thereby determining the duration, quantity and other information of the preset time period.

[0144] Exemplarily, the department category may include obstetrics and gynecology and orthopedics, and the duration of the preset period determined according to obstetrics and gynecology is shorter than the duration of the preset period determined according to orthopedics.

[0145] Obstetrics and gynecology usually need to check fetal related information, such as different parts of the fetus, so it is necessary to frequently switch the ultrasound scanning range, and the probe stays in the same part for a short time. Therefore, the duration of the preset time period determined by obstetrics and gynecology can be shorter. In contrast, orthopedics usually need to maintain the same ultrasound scanning range for a long time to understand the state of bones and muscles, and the probe stays in the same part for a longer time, so the duration of the preset time period determined by orthopedics will also be longer.

[0146] For example, for a 1-minute ultrasound video, obstetrics and gynecology can be divided into 6 preset time periods, and orthopedics can be divided into 3 preset time periods.

[0147] In the above technical solution, the preset time period is determined according to the type of department used to examine the tissue. The preset time period can be determined in combination with the ultrasound examination habits of the medical staff of the department to better meet the needs of users.

[0148] Fig.11 A schematic diagram of an ultrasound image segmentation method according to an embodiment of the present invention is shown.

[0149] like Fig.11As shown, for an ultrasound video composed of 1800 ultrasound image frames collected at 30 frames per second, the corresponding duration of this ultrasound video is 60 s. This ultrasound video can be an ultrasound video related to musculoskeletal, such as an ultrasound video related to the hand. In this ultrasound video, there can be muscle tissues and bones related to the hand, such as the extensor digiti minimi tendon, extensor digitorum indicis tendon, radius, etc., a total of 10 tissues. One preset period can be determined according to the acquisition moment corresponding to every 300 frames, and 6 preset periods are obtained: 0 s < t <= 10 s, 10 s < t <= 20 s, 20 s < t <= 30 s, 30 s < t <= 40 s, 40 s < t <= 50 s, and 50 s < t <= 60 s. There are 10 tissues in the 1800-frame ultrasound video. For these 10 tissues and 6 preset periods, 60 different initial query vectors query can be determined. For the 10 tissues in any one preset period, the 10 initial query vectors query corresponding to the 10 tissues respectively in this preset period can be used as the initial query vectors q corresponding to the 10 tissues respectively in this preset period. Among them, for any one of the 10 tissues, when 0 s < t <= 10 s, q = q0; when 10 s < t <= 20 s, q = q1; when 20 s < t <= 30 s, q = q2; when 30 s < t <= 40 s, q = q3; when 40 s < t <= 50 s, q = q4; when 50 s < t <= 60 s, q = q5. Among them, t represents the moment in different periods. Among them, q0, q1, q2, q3, q4, q5 represent the initial query vectors corresponding to this tissue in different periods.

[0150] Then each image frame in this ultrasound video can be used as a target ultrasound image respectively, and the target ultrasound image and the target query vector corresponding to the target tissue in the target ultrasound image are input into an image segmentation model to obtain a segmentation image of the ultrasound video with musculoskeletal segmented out.

[0151] Exemplarily, this ultrasound video can also be directly input into the image segmentation model to determine the initial query vector, calculate the target query vector according to the steps in the above ultrasound image segmentation method, and then perform image segmentation to output a segmentation image of this ultrasound video with musculoskeletal segmented out. The musculoskeletal as the segmentation object is the tissue expected to be segmented among these 10 tissues, that is, the target tissue.

[0152] Then the segmentation effect of the segmentation image of this ultrasound video can be evaluated, and the segmentation effect of the segmentation image of this ultrasound video is shown in Table 1 below:

[0153] Table 1

[0154] Image segmentation method Query vector number AP Value Split method 1 1 6.2 Split method 2 1800 10.2 Split method 3 10 17.9 Split method 4 18000 17.5 Split method 5 60 19.1

[0155] Among them, the AP value is used to evaluate the segmentation effect of the segmentation image, and the AP value of segmentation method 5 represents according to the above Fig.11 The segmentation effect of the segmented image obtained by the ultrasound image segmentation method shown. In segmentation method 1, the query vectors corresponding to the 1800 ultrasound image frames are the same, and there is only one. In segmentation method 2, the query vectors corresponding to the 1800 ultrasound image frames are different, but the target tissue in each ultrasound image frame is taken as a single target instance corresponding to the query vector corresponding to the ultrasound image frame. In segmentation method 3, 10 target tissues in 1800 ultrasound image frames are taken as target instances, and 10 query vectors corresponding to the 1800 ultrasound image frames are generated at the same time as the query vectors corresponding to the 1800 ultrasound image frames. In segmentation method 4, each target tissue in each frame of the 1800 ultrasound image frames is taken as a single target instance, and a total of 18000 query vectors of 1800*10 are obtained. As can be seen from Table 1, Fig.11 The ultrasound image segmentation method shown in FIG.

[0156] The target tissue may have the following morphology: Fig.11 The ultrasonic image frames in the six preset time periods shown in the figure gradually change. Therefore, the reference time of each preset time period can be determined first, and then the target query vector corresponding to the target tissue in the ultrasonic image frame as the target ultrasonic image can be determined according to the reference time, the time corresponding to the ultrasonic image frame as the target ultrasonic image, and the above-mentioned ultrasonic image segmentation method. When the reference time is 5s, 15s, 25s, 35s, 45s, and 55s, the target query vector Q corresponding to any target tissue in the target ultrasonic image can be determined according to the above formula 1 and the following table 2:

[0157] Table 2

[0158] Ultrasound image corresponding to the time The value of the target query vector Q 0s <t<=5s q0 5s<t<=15s q0*∣15-t∣ / 10+q1*∣5-t∣ / 10 15s<t<=25s q1*∣25-t∣ / 10+q2*∣15-t∣ / 10 <h2 style=";text-align:left;direction:ltr">25s <t<=35s <h2 style=";text-align:left;direction:ltr"> q2*∣35-t∣ / 10+q3*∣25-t∣ / 10 35s <t<=45s q3*∣45-t∣ / 10+q4*∣35-t∣ / 10 45s<t<=55s q4*∣55-t∣ / 10+q5*∣45-t∣ / 10 t>=55s q5

[0159] When the target query vector Q in Table 2 and the image segmentation model in the above ultrasound image segmentation method are used to segment the ultrasound video, the AP value of the segmented image is 21.0, which is better than the above segmentation methods 1-4 and achieves the optimal segmentation effect.

[0160] Fig.12 FIG. 4 is a schematic block diagram of an ultrasonic image segmentation device according to an embodiment of the present invention. Fig.12 As shown, the ultrasound image segmentation device includes a receiving module 1210 , a determining module 1220 , and a segmentation module 1230 .

[0161] The receiving module 1210 is used to acquire an ultrasound image sequence, where the ultrasound image sequence includes a plurality of ultrasound images at different moments, and the plurality of ultrasound images include ultrasound images at a plurality of preset time periods.

[0162] The determination module 1220 is used for determining at least one target ultrasound image in the ultrasound image sequence.

[0163] Based on an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located, a target query vector corresponding to the target tissue in the target ultrasound image is determined, wherein the initial query vector is determined based on features of the tissue in the ultrasound image sequence in multiple ultrasound images in the corresponding preset time period.

[0164] The segmentation module 1230 is used to perform image segmentation on the target ultrasound image according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image.

[0165] Exemplarily, the ultrasound image segmentation device further includes a tissue determination submodule and a first determination submodule. The tissue determination submodule is used to determine all tissues included in the ultrasound image sequence. The first determination submodule is used to determine the initial query vector corresponding to each tissue in each preset time period based on the features of all ultrasound images of each tissue in each preset time period.

[0166] Exemplarily, the moment corresponding to the target ultrasound image is the first moment, the preset time period in which the first moment is located is the first preset time period, and the determination module 1220 includes a reference moment determination submodule, a first target time period determination submodule and a query vector calculation submodule. The reference moment determination submodule is used to determine a moment in the preset time period as a reference moment for each preset time period. The first target time period determination submodule is used to determine the first target time period in multiple preset time periods according to the reference moment and the first moment of the first preset time period, wherein the first target time period includes the first preset time period and a preset time period adjacent to the first preset time period, and the adjacent preset time period is closer to the first moment than to the reference moment of the first preset time period. The query vector calculation submodule is used to calculate the target query vector corresponding to the target tissue in the target ultrasound image according to the initial query vector, the reference moment and the first moment corresponding to each tissue in the first target time period.

[0167] Exemplarily, the query vector calculation submodule includes a first calculation submodule, a second calculation submodule and a third calculation submodule. The first calculation submodule is used to determine, for any target tissue in the target ultrasound image, a first coefficient of an initial query vector corresponding to the target tissue in the first preset time period according to a first time interval between a reference moment and the first moment in adjacent preset time periods. The second calculation submodule is used to determine a second coefficient of an initial query vector corresponding to the target tissue in an adjacent preset time period according to a second time interval between a reference moment and the first moment in the first preset time period. The third calculation submodule is used to determine a final query vector as a target query vector corresponding to the target tissue in the target ultrasound image according to the first coefficient, the second coefficient, the initial query vector corresponding to the target tissue in the first preset time period and the initial query vector corresponding to the target tissue in the adjacent preset time period.

[0168] Exemplarily, any reference moment determined is a middle moment of a preset time period in which the reference moment is located.

[0169] Exemplarily, the ultrasonic image segmentation device further includes a duration division submodule and a time period division submodule. The duration division submodule is used to determine the division duration of the ultrasonic image sequence according to the inspection item category of the ultrasonic examination and / or the inspection duration corresponding to the inspection item category, and the inspection item category corresponds to the tissue in the ultrasonic image sequence; the time period division submodule is used to divide multiple ultrasonic images at different times according to the division duration corresponding to the acquired ultrasonic image sequence after acquiring the ultrasonic image sequence, so as to obtain ultrasonic images of multiple preset time periods.

[0170] Exemplarily, the duration division submodule includes a first division submodule and a second division submodule.

[0171] The first division submodule is used to determine the division time of the ultrasound image sequence as the first division time if the examination item category is an examination item corresponding to the heart. The second division submodule is used to determine the division time of the ultrasound image sequence as the second division time if the examination item category is an examination item corresponding to obstetrics and gynecology, an examination item corresponding to the thyroid gland, an examination item corresponding to the breast, an examination item corresponding to the musculoskeletal shoulder joint, an examination item corresponding to the musculoskeletal ankle joint, or an examination item corresponding to the abdominal organ. The first division time is greater than the second division time.

[0172] Exemplarily, the third target time is a middle time in the third preset time period.

[0173] Exemplarily, the duration of the preset time period is less than or equal to the set value.

[0174] Exemplarily, the target tissue includes muscle tissue and / or bone.

[0175] According to another aspect of the present invention, an electronic device is provided. Fig.13 FIG. 1 shows a schematic block diagram of an electronic device according to an embodiment of the present invention. Fig.13 As shown, the electronic device includes a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions are used to execute the ultrasound image segmentation method as described above when the processor is executed.

[0176] In addition, according to another aspect of the present invention, a storage medium is also provided, on which program instructions are stored, and when the program instructions are executed by a computer or a processor, the computer or the processor performs the corresponding steps of the above-mentioned ultrasound image segmentation method according to the embodiment of the present invention, and is used to implement the corresponding modules in the above-mentioned ultrasound image segmentation device according to the embodiment of the present invention. The storage medium may include, for example, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disk read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.

[0177] According to another aspect of the present invention, a computer program product is provided, comprising computer program instructions, wherein the computer program instructions are used to execute the above-mentioned ultrasound image segmentation method when running.

[0178] A person skilled in the art can understand the specific implementation and beneficial effects of the above-mentioned ultrasonic image segmentation device, electronic device, storage medium and computer program product by reading the above-mentioned detailed description of the ultrasonic image segmentation method, which will not be described here for the sake of brevity.

[0179] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present invention thereto. Various changes and modifications may be made therein by one of ordinary skill in the art without departing from the scope and spirit of the present invention. All such changes and modifications are intended to be included within the scope of the present invention as required by the appended claims.

[0180] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0181] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0182] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0183] Similarly, it should be understood that in order to streamline the present invention and help understand one or more of the various inventive aspects, in the description of the exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the method of the present invention should not be interpreted as reflecting the following intention: the claimed invention requires more features than the features explicitly stated in each claim. More specifically, as reflected in the corresponding claims, the inventive point is that the corresponding technical problem can be solved with less than all the features of a single disclosed embodiment. Therefore, the claims following the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present invention.

[0184] It will be understood by those skilled in the art that, except for mutually exclusive features, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this specification may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature that provides the same, equivalent or similar purpose.

[0185] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims, any one of the claimed embodiments may be used in any combination.

[0186] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some modules in an ultrasonic image segmentation device according to an embodiment of the present invention. The present invention may also be implemented as a device program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0187] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

[0188] The above is only a specific embodiment of the present invention or an explanation of a specific embodiment. The protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for segmenting an ultrasonic image, characterized in that: The method comprises: Acquire an ultrasound image sequence, wherein the ultrasound image sequence includes a plurality of ultrasound images at different moments, and the plurality of ultrasound images include ultrasound images at a plurality of preset time periods; For at least one target ultrasound image in the ultrasound image sequence, Determine a target query vector corresponding to a target tissue in the target ultrasound image according to an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on features of the tissue in the ultrasound image sequence in a plurality of ultrasound images in a corresponding preset time period; The target ultrasound image is segmented according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image.

2. The method according to claim 1, characterized in that The method further comprises: determining all tissues included in the ultrasound image sequence; Based on the features of each tissue in all ultrasound images in each preset time period, an initial query vector corresponding to each tissue in each preset time period is determined.

3. The method according to claim 2, characterized in that The moment corresponding to the target ultrasonic image is a first moment, and the preset time period in which the first moment is located is a first preset time period; The step of determining a target query vector corresponding to a target tissue in the target ultrasound image according to an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located comprises: For each preset time period, determining a moment in the preset time period as a reference moment; Determine a first target period among the plurality of preset periods according to a reference time of the first preset period and the first time, wherein the first target period includes the first preset period and a preset period adjacent to the first preset period, and the adjacent preset period is closer to the first time than to the reference time of the first preset period; A target query vector corresponding to the target tissue in the target ultrasound image is calculated according to the initial query vectors respectively corresponding to each tissue in the first target time period, the reference time and the first time.

4. The method according to claim 3, characterized in that The step of calculating a target query vector corresponding to a target tissue in the target ultrasound image according to the initial query vectors corresponding to each tissue in the first target time period, the reference time and the first time comprises: For any target tissue in the target ultrasound image, Determine a first coefficient of an initial query vector corresponding to the target tissue in the first preset time period according to a first time interval between a reference time in the adjacent preset time period and the first time period; Determine, according to a second time interval between a reference moment in the first preset time period and the first moment, a second coefficient of an initial query vector corresponding to the target tissue in the adjacent preset time period; A final query vector is determined as a target query vector corresponding to the target tissue in the target ultrasound image based on the first coefficient, the second coefficient, an initial query vector corresponding to the target tissue in the first preset time period, and an initial query vector corresponding to the target tissue in the adjacent preset time period.

5. The method according to claim 3, characterized in that: Any determined reference time is the middle time of the preset time period in which the reference time is located.

6. The method according to claim 1, characterized in that The method further comprises: Determining the division duration of the ultrasound image sequence according to the examination item category of the ultrasound examination and / or the examination duration corresponding to the examination item category, the examination item category corresponding to the tissue in the ultrasound image sequence; Accordingly, after acquiring the ultrasound image sequence, the method further includes: The multiple ultrasound images at different moments are divided according to the division time lengths corresponding to the acquired ultrasound image sequence to obtain the ultrasound images of the multiple preset time periods.

7. The method according to claim 6, characterized in that The determining of the division duration of the ultrasound image sequence according to the ultrasound examination item category and / or the examination duration corresponding to the examination item category includes: If the examination item category is an examination item corresponding to the heart, the division time length of the ultrasound image sequence is determined as the first division time length; If the examination item category is an examination item corresponding to obstetrics and gynecology, an examination item corresponding to the thyroid gland, an examination item corresponding to the breast, an examination item corresponding to the musculoskeletal shoulder joint, an examination item corresponding to the musculoskeletal ankle joint, or an examination item corresponding to the abdominal organ, the division time length of the ultrasound image sequence is determined to be the second division time length; The first division duration is greater than the second division duration.

8. The method according to any one of claims 1 to 7, characterized in that: The duration of the preset time period is less than or equal to the set value.

9. The method according to any one of claims 1 to 7, characterized in that: The target tissue includes muscle tissue and / or bone.

10. An ultrasonic image segmentation device, characterized in that: include: A receiving module, used for acquiring an ultrasound image sequence, wherein the ultrasound image sequence includes a plurality of ultrasound images at different moments, and the plurality of ultrasound images include ultrasound images at a plurality of preset time periods; a determination module, for at least one target ultrasound image in the ultrasound image sequence, determining a target query vector corresponding to a target tissue in the target ultrasound image according to an initial query vector corresponding to the tissue in the ultrasound image sequence and a preset time period in which the target ultrasound image is located, wherein the initial query vector is determined based on features of a plurality of ultrasound images of the tissue in the ultrasound image sequence in a corresponding preset time period; The segmentation module is used to perform image segmentation on the target ultrasound image according to the target query vector to obtain a segmented image of the target tissue in the target ultrasound image.

11. An electronic device comprising a processor and a memory, characterized in that: The memory stores computer program instructions, which are used by the processor to execute the ultrasound image segmentation method according to any one of claims 1 to 9 when executed.

12. A storage medium having program instructions stored thereon, characterized in that: The program instructions are used to execute the ultrasound image segmentation method according to any one of claims 1 to 9 when running.

13. A computer program product comprising computer program instructions, characterized in that The computer program instructions are used to execute the ultrasound image segmentation method according to any one of claims 1 to 9 when executed.

Images