Volume measurement method and device of target tissue and storage medium

Abstract

The invention relates to a volume measurement method and device of a target tissue and a storage medium. The method comprises the following steps: in response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue; adjusting the plurality of sub-contours of the target tissue, and performing real-time correction processing on the plurality of adjusted sub-contours according to an online trained classification model to obtain a tissue contour of the target tissue; and determining the volume of the target tissue according to the tissue contour of the target tissue. By adopting the method, the accuracy of volume measurement can be improved.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular to a method, device and storage medium for measuring the volume of a target tissue. Background Art

[0002] In the medical field, the measurement of tissue volume is of great significance to the patient's health examination and subsequent diagnosis and treatment, and is related to the focus of the entire diagnosis and treatment.

[0003] In the related art, when the volume measurement of the target tissue is required, the target tissue is usually segmented by an AI algorithm, and then the doctor manually outlines multiple sub-contours of the target tissue, and then directly generates a complete contour based on the outlined multiple sub-contours, and calculates the volume of the target tissue based on the complete contour.

[0004] However, in the related art, when measuring the volume of the target tissue, due to the low quality of ultrasonic imaging, the existence of noise and artifacts, etc., the robustness of the AI ​​algorithm segmentation effect is poor, which in turn affects the accuracy of the volume calculation. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device and storage medium for measuring the volume of a target tissue in order to address the above technical issues, so as to improve the accuracy of volume measurement.

[0006] In a first aspect, the present application provides a method for measuring the volume of a target tissue, comprising:

[0007] In response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue;

[0008] Adjusting multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0009] According to the tissue contour of the target tissue, the volume of the target tissue is determined.

[0010] In one embodiment, obtaining a plurality of sub-contours of the target tissue according to the three-dimensional image of the target tissue includes:

[0011] Inputting the three-dimensional image into a preset segmentation model to obtain a candidate segmentation result of the target tissue output by the segmentation model;

[0012] Based on the candidate segmentation results, multiple sub-contours of the target tissue are determined.

[0013] In one embodiment, a plurality of sub-contours of the target tissue are determined based on the candidate segmentation results, including:

[0014] According to the candidate segmentation result, obtaining the rotation axis of the candidate segmentation result;

[0015] According to the rotation axis and the preset angle step, multiple rotation surfaces of the candidate segmentation results are obtained;

[0016] A plurality of sub-contours of the target tissue are determined according to a plurality of rotation planes of the candidate segmentation results.

[0017] In one embodiment, obtaining a rotation axis of the candidate segmentation result according to the candidate segmentation result includes:

[0018] Get the maximum circumscribed cube of the candidate segmentation result;

[0019] Expand the largest circumscribed cube outward by multiple pixels to obtain the bounding box of the candidate segmentation result;

[0020] The rotation axis of the candidate segmentation result is determined based on the multiple volume diagonals of the bounding box.

[0021] In one embodiment, obtaining multiple rotation planes of the candidate segmentation results according to the rotation axis and the preset angle step size includes:

[0022] Obtain multiple vectors perpendicular to the rotation axis, and determine any one of the vectors perpendicular to the rotation axis as the initial normal vector;

[0023] Based on the angle step, starting from the initial normal vector, multiple candidate normal vectors are obtained;

[0024] According to the rotation axis, the rotation center, the initial normal vector and each candidate normal vector, a plurality of rotation planes of the candidate segmentation result are determined.

[0025] In one embodiment, multiple sub-profiles of the target tissue are adjusted, including:

[0026] Displaying a contour adjustment interface; the contour adjustment interface includes multiple sub-contours of the target tissue;

[0027] Receiving adjustment position information of a user operating on a plurality of sub-contours;

[0028] The multiple sub-contours are adjusted according to the adjustment position information of the multiple sub-contours to generate multiple adjusted sub-contours.

[0029] In one embodiment, the adjusted multiple sub-profiles are corrected in real time according to the online trained classification model to obtain the tissue profile of the target tissue, including:

[0030] Get the pixel coordinates of multiple pixels in the bounding box;

[0031] Input the pixel coordinates of each pixel into the classification model to obtain the classification results of each pixel;

[0032] According to the classification results of each pixel point, the adjusted multiple sub-contours are corrected in real time to obtain the tissue contour of the target tissue.

[0033] In one embodiment, the method further comprises:

[0034] Obtaining contour coordinate data of the adjusted multiple sub-contours;

[0035] Normalizing each contour coordinate data to obtain normalized contour coordinate data of a plurality of adjusted sub-contours;

[0036] According to each normalized contour coordinate data, the preset neural network is trained online to obtain a classification model.

[0037] In a second aspect, the present application also provides a device for measuring the volume of a target tissue, comprising:

[0038] A sub-contour acquisition module, configured to acquire a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue in response to a volume measurement request of the target tissue;

[0039] A first determination module is used to adjust multiple sub-contours of the target tissue, and perform real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0040] The second determination module is used to determine the volume of the target tissue according to the tissue contour of the target tissue.

[0041] In a third aspect, the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in any one of the embodiments of the first aspect when executing the computer program.

[0042] In a fourth aspect, the embodiments of the present application further provide a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps in any one of the embodiments of the first aspect are implemented.

[0043] In a fifth aspect, the embodiments of the present application further provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps in any one of the embodiments in the first aspect are implemented.

[0044] The above-mentioned target tissue volume measurement method, device and storage medium, by responding to the volume measurement request of the target tissue, obtains multiple sub-contours of the target tissue according to the three-dimensional image of the target tissue, and then adjusts the multiple sub-contours of the target tissue, and performs real-time correction processing on the adjusted multiple sub-contours according to the online classification model to obtain the tissue contour of the target tissue, and finally determines the volume of the target tissue according to the tissue contour of the target tissue. In this method, when it is necessary to measure the volume of the target tissue, multiple sub-contours of the target tissue are obtained according to the three-dimensional image of the target tissue, and then before generating the three-dimensional tissue contour of the target tissue according to the multiple sub-contours, the multiple sub-contours are adjusted and the adjusted multiple sub-contours are corrected in real time according to the online trained classification model, so that the multiple sub-contours of the target tissue are more accurate, and the obtained three-dimensional tissue contour of the target tissue will also be more accurate, so that the volume of the target tissue measured is also more accurate, and the accuracy of volume measurement is improved. In addition, when obtaining multiple sub-contours of the target tissue, the method is not manually outlined by the doctor, but extracted by a computer device, which reduces the workload of the doctor in this process. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 is an internal structure diagram of a computer device in one embodiment;

[0047] Figure 2 is a schematic flow chart of a method for measuring the volume of a target tissue in one embodiment;

[0048] Figure 3 A schematic diagram of a process of obtaining multiple sub-profiles in one embodiment;

[0049] Figure 4 A schematic diagram of a process of obtaining multiple sub-profiles in another embodiment;

[0050] Figure 5 A schematic diagram of a process of obtaining a rotating shaft in one embodiment;

[0051] Figure 6 A schematic diagram of a process for obtaining multiple rotating surfaces in one embodiment;

[0052] Figure 7 is a schematic diagram of a process of fine-tuning multiple sub-profiles in one embodiment;

[0053] Figure 8 A schematic diagram of a process for determining an organization profile in one embodiment;

[0054] Fig. 9 A schematic diagram of a process for building a classification model in one embodiment;

[0055] Fig.10 A schematic flow chart of a method for measuring the volume of a target tissue in another embodiment;

[0056] Fig.11 FIG. 4 is a schematic diagram of the structure of a device for measuring the volume of a target tissue in an embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0058] The target tissue volume measurement method provided in the embodiment of the present application can be applied to a computer device. The computer device can be a server, and its internal structure diagram can be as follows: Figure 1 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for measuring the volume of a target tissue is implemented. Those skilled in the art can understand that Figure 1 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0059] In the medical field, the measurement of tissue volume is of great significance to the patient's health examination and subsequent diagnosis and treatment, and is related to the focus of the entire diagnosis and treatment.

[0060] Taking three-dimensional ultrasound volume measurement as an example, in the related technology, when the volume of the target tissue needs to be measured, the doctor is usually required to manually select the rotation axis and outline the contour on each angle plane, and then directly generate a complete contour based on the multiple outlined contours, and then calculate the volume of the target tissue based on the complete contour. However, as the number of angle planes increases, doctors need to spend more time and energy to complete this task. Then when the number of angle planes is small, the lack of sufficient surface information will lead to poor 3D reconstruction effects, which will ultimately affect the accuracy of volume measurement. And the volume calculation directly based on the manually outlined contour will also have the problem of low volume measurement accuracy.

[0061] Based on this, the present application provides a method for measuring the volume of a target tissue. When the volume of the target tissue needs to be measured, multiple sub-contours of the target tissue are obtained based on the three-dimensional image of the target tissue, and then before generating the three-dimensional tissue contour of the target tissue based on the multiple sub-contours, the multiple sub-contours are fine-tuned and the fine-tuned multiple sub-contours are corrected according to the classification model, so that the multiple sub-contours of the target tissue are more accurate, and the obtained three-dimensional tissue contour of the target tissue will also be more accurate, so that the volume of the target tissue measured is also more accurate, and the accuracy of the volume measurement is improved. In addition, when obtaining multiple sub-contours of the target tissue, the method is not manually outlined by the doctor, but the multiple sub-contours are extracted by a computer device, which reduces the workload of the doctor in the process.

[0062] It should be noted that the beneficial effects brought about by the embodiments of the present application or the technical problems solved are not limited to this one, but may also include other implicit or related problems. For details, please refer to the description of the following embodiments.

[0063] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0064] In an exemplary embodiment, Figure 2 As shown, a method for measuring the volume of a target tissue is provided, and the method is applied to Figure 1 The computer device in the example is used to illustrate, including the following steps 201 to 203. Among them:

[0065] S201 , in response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue.

[0066] The target tissue refers to any tissue that requires volume measurement, such as the heart, liver, gallbladder, kidney, prostate, thyroid, etc.

[0067] The volume measurement request of the target tissue refers to a request triggered when the volume measurement of the target tissue needs to be performed in a volume measurement application of the computer device. Naturally, after receiving the volume measurement request, the computer device needs to perform volume measurement on the target tissue.

[0068] In one embodiment, the request is triggered by a user in a volume measurement application through a computer device interface; wherein the triggering method includes but is not limited to external input device clicks, voice control, touch screen triggering, etc., and the embodiment of the present application does not limit the triggering method.

[0069] In another embodiment, the request is triggered according to a preset automatic program; for example, if the computer device receives the three-dimensional image of the target tissue transmitted by the image acquisition device, the volume measurement request of the target tissue is automatically triggered according to the preset program; or, after the computer device receives the three-dimensional image of the target tissue transmitted by the image acquisition device, if the volume measurement application is in an idle state, the volume measurement request of the target tissue is automatically triggered according to the preset program. The embodiment of the present application does not limit the triggering method of the volume measurement request.

[0070] After receiving the volume measurement request of the target tissue, the computer device starts to measure the volume of the target tissue, firstly obtains a three-dimensional image of the target tissue, and then obtains multiple sub-contours of the target tissue based on the three-dimensional image of the target tissue.

[0071] Among them, the three-dimensional image of the target tissue can be carried in the volume measurement request of the target tissue, that is, when the volume measurement request of the target tissue is triggered, the three-dimensional image of the target tissue is uploaded to the computer device at the same time. In this way, when the computer device receives the volume measurement request of the target tissue, it can obtain the three-dimensional image of the target tissue carried in the request.

[0072] It should be noted that the three-dimensional image is obtained by scanning the target tissue using an image acquisition device, wherein the image acquisition device may include but is not limited to an ultrasonic imaging device, a magnetic resonance imaging device, an electronic computer tomography device, etc.

[0073] When acquiring multiple sub-contours of a target tissue, the target tissue can be segmented from a three-dimensional image of the target tissue, and then divided into multiple two-dimensional sections, and then contour extraction is performed on these two-dimensional sections. For example, contour extraction is performed on the two-dimensional sections in sequence to obtain multiple sub-contours of the target tissue.

[0074] Taking the target tissue as the liver and the three-dimensional image as an ultrasound image as an example, when there is a need to measure the volume of the liver, the user can trigger a request for the volume measurement of the liver on the computer device interface. After the computer device receives the request for the volume measurement of the liver, it first segments the liver ultrasound image carried in the volume measurement request to obtain the liver, and then divides the liver into multiple two-dimensional sections, and extracts the contour of each two-dimensional section to obtain multiple sub-contours of the liver.

[0075] S202, adjusting multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue.

[0076] After obtaining multiple sub-contours of the target tissue, a three-dimensional tissue contour of the target tissue can be fitted based on the multiple sub-contours. In the embodiment of the present application, in order to improve the accuracy of the three-dimensional tissue contour and thus improve the accuracy of the volume measurement, after obtaining multiple sub-contours of the multiple target tissues, the multiple sub-contours are adjusted multiple times to make the multiple sub-contours of the target tissue more accurate, the tissue contour fitted according to the multiple sub-contours more accurate, and thus the measured volume of the target tissue is also more accurate.

[0077] In an embodiment of the present application, the multiple sub-contours of the target tissue are first adjusted, and then the adjusted multiple sub-contours are corrected again to obtain multiple corrected sub-contours, and then the three-dimensional contour of the target tissue is fitted based on the multiple corrected sub-contours.

[0078] Among them, the adjustment of multiple sub-contours of the target tissue can be performed by manually inputting each sub-contour, or by automatically adjusting each sub-contour using a professional knowledge base, wherein the professional knowledge base can include a universal template contour, and the sub-contour is adjusted through the template contour.

[0079] Furthermore, after adjusting the multiple sub-contours of the target tissue, the adjusted multiple sub-contours are corrected in real time using the online trained classification model to obtain the tissue contour of the target tissue. The real-time correction of the sub-contours in the embodiment of the present application is achieved because the classification model used is an online trained classification model rather than a pre-trained classification model.

[0080] Among them, the online training classification model updates the weight after a piece of data is trained. When new data is generated, the weight and bias parameters of the model are automatically updated in real time. By continuously accepting new data, the accuracy of the model is improved, and the adaptability and adjustability to data are stronger.

[0081] S203, determining the volume of the target tissue according to the tissue contour of the target tissue.

[0082] Exemplarily, the number of all pixels belonging to the target tissue can be obtained according to the tissue contour of the target tissue to obtain the pixel volume of the target tissue, and then the pixel volume of the target tissue can be converted into the actual physical volume of the target tissue through the conversion relationship between the pixel distance and the actual physical distance, thereby obtaining the volume of the target tissue, wherein the conversion relationship can be: physical size = physical resolution per unit pixel * number of unit pixels.

[0083] In the target tissue volume measurement method provided by the embodiment of the present application, by responding to the volume measurement request of the target tissue, according to the three-dimensional image of the target tissue, multiple sub-contours of the target tissue are obtained, and then the multiple sub-contours of the target tissue are adjusted, and the adjusted multiple sub-contours are corrected in real time according to the online training classification model to obtain the tissue contour of the target tissue, and finally the volume of the target tissue is determined according to the tissue contour of the target tissue. In this method, when the volume measurement of the target tissue is required, multiple sub-contours of the target tissue are obtained according to the three-dimensional image of the target tissue, and then before the three-dimensional tissue contour of the target tissue is generated according to the multiple sub-contours, the multiple sub-contours are adjusted and the adjusted multiple sub-contours are corrected according to the online training classification model, so that the multiple sub-contours of the target tissue are more accurate, and then the obtained three-dimensional tissue contour of the target tissue will also be more accurate, so that the volume of the target tissue measured is also more accurate, and the accuracy of the volume measurement is improved. In addition, when the method obtains the multiple sub-contours of the target tissue, it is not manually outlined by the doctor, but the multiple sub-contours are extracted by the computer device, which reduces the workload of the doctor in this process.

[0084] Based on the above embodiment, an embodiment is provided to illustrate the process of acquiring multiple sub-contours of the target tissue according to the three-dimensional image of the target tissue.

[0085] In an exemplary embodiment, Figure 3 As shown, this embodiment includes the following steps 301 to 302:

[0086] S301, inputting a three-dimensional image into a preset segmentation model to obtain a candidate segmentation result of a target tissue output by the segmentation model.

[0087] When acquiring multiple sub-contours of the target tissue, the target tissue is first roughly segmented to roughly locate the position of the target tissue in the three-dimensional image.

[0088] Exemplarily, the computer device inputs the three-dimensional image of the target tissue into a pre-trained segmentation model, and the segmentation model processes the three-dimensional image of the target tissue and outputs a rough segmentation result of the target tissue. Naturally, the rough segmentation result is the candidate segmentation result of the target tissue. Before the three-dimensional image of the target tissue is input into the segmentation model, the three-dimensional image of the target tissue is pre-processed and normalized to [0, 1].

[0089] The segmentation model can be obtained by training a preset neural network using a small amount of labeled image data, wherein the preset neural network can be a convolutional neural network, a recurrent neural network, a deep neural network, etc. It should be noted that when training the segmentation model, the image input to the neural network is also a normalized image, which can accelerate the convergence of the training network on the one hand, and eliminate the adverse effects caused by singular sample data on the other hand, thereby improving the prediction accuracy of the model.

[0090] S302: Determine multiple sub-contours of the target tissue according to the candidate segmentation results.

[0091] After obtaining the candidate segmentation results of the target tissue, the candidate segmentation results can be divided according to preset rules to obtain multiple two-dimensional sections. Then, a preset contour extraction algorithm can be used to perform contour extraction on the multiple two-dimensional sections respectively to obtain multiple contours. Naturally, the multiple contours are multiple sub-contours of the target tissue.

[0092] In the volume measurement method of the target tissue provided in the embodiment of the present application, by inputting the three-dimensional image into the preset segmentation model, the candidate segmentation result of the target tissue output by the segmentation model is obtained, and then multiple sub-contours of the target tissue are determined according to the candidate segmentation results. In this method, the target tissue is processed by the segmentation model to obtain the candidate segmentation result of the target tissue, and the position of the target tissue in the three-dimensional image can be roughly located. Based on this, multiple sub-contours of the target tissue can be extracted from the candidate segmentation results, providing an optional way to quickly obtain sub-contours.

[0093] Based on the above embodiment, an embodiment is provided to illustrate the process of determining multiple sub-contours of the target tissue according to the candidate segmentation results.

[0094] In an exemplary embodiment, Figure 4 As shown, this embodiment includes the following steps 401 to 403:

[0095] S401, obtaining a rotation axis of the candidate segmentation result according to the candidate segmentation result.

[0096] When acquiring multiple sub-contours of a target tissue, a rotation axis may be determined first, so as to obtain multiple two-dimensional sections, ie, rotation planes, according to the rotation axis, and then the sub-contours may be extracted based on the multiple rotation planes.

[0097] In one embodiment, the rotation axis can be directly given by the computer device; for example, the volume measurement application in the computer device pre-sets a rotation axis acquisition procedure. When the rotation axis needs to be acquired, the computer device automatically acquires the rotation axis according to the preset procedure and can be displayed in the interface of the computer device. The interface also displays candidate segmentation results, and the rotation axis is displayed at the corresponding position in the candidate segmentation results.

[0098] In another embodiment, the rotation axis can be determined by the user from multiple rotation axes provided by the computer device; for example, a rotation axis component module is provided in the interface of the computer device, and when the user clicks on the rotation axis component module, multiple rotation axes can be displayed, and the user can select one rotation axis from them as a candidate segmentation result.

[0099] S402: Acquire multiple rotation surfaces of the candidate segmentation results according to the rotation axis and the preset angle step.

[0100] After obtaining the rotation axis of the candidate segmentation result, the computer device can segment the candidate segmentation result according to the rotation axis and a preset angle step to obtain multiple two-dimensional sections, that is, multiple rotation planes of the candidate segmentation result.

[0101] Exemplarily, the computer device may first obtain a preset segmentation plane, start from the preset segmentation plane, rotate with the rotation axis as the axis and according to the preset angle step, and segment once every rotation of one angle step to obtain a two-dimensional section from the candidate segmentation result, until the total rotation angle is greater than or equal to 180°, and then stop, and determine all the two-dimensional sections obtained after segmentation as multiple rotation planes of the candidate segmentation result. The preset angle step can be set according to actual application requirements, for example, it can be 15° as the angle step.

[0102] S403, determining multiple sub-contours of the target tissue according to multiple rotation planes of the candidate segmentation results.

[0103] After obtaining a plurality of rotation planes of the candidate segmentation results, a preset contour extraction algorithm may be used to perform contour extraction on the plurality of rotation planes to obtain a plurality of sub-contours of the target tissue.

[0104] In the volume measurement method of the target tissue provided in the embodiment of the present application, the rotation axis of the candidate segmentation result is obtained according to the candidate segmentation result, and multiple rotation planes of the candidate segmentation result are obtained according to the rotation axis and the preset angle step, and multiple sub-contours of the target tissue are determined according to the multiple rotation planes of the candidate segmentation result. In this method, by obtaining the rotation axis of the candidate segmentation result, the candidate segmentation result can be divided by the rotation axis to obtain multiple rotation planes, so that multiple sub-contours can be extracted from the multiple rotation planes, which provides an optional way to quickly determine the sub-contours.

[0105] Based on the above embodiment, an embodiment is provided to illustrate the process of obtaining the rotation axis of the candidate segmentation result according to the candidate segmentation result.

[0106] In an exemplary embodiment, Figure 5 As shown, this embodiment includes the following steps 501 to 503:

[0107] S501, obtaining the largest circumscribed cube of the candidate segmentation result.

[0108] S502: Expand the largest circumscribed cube outward by multiple pixels to obtain a bounding box of the candidate segmentation result.

[0109] When obtaining the rotation axis of the candidate segmentation result, the maximum circumscribed cube of the candidate segmentation result is obtained to obtain the rotation axis according to the maximum circumscribed cube.

[0110] The computer device draws the maximum circumscribed cube of the candidate segmentation result based on the candidate segmentation result, and then expands the maximum circumscribed cube of the candidate segmentation result outward by n pixels to obtain a new cube. Naturally, the new cube is the bounding box of the candidate segmentation result. For example, after obtaining the maximum circumscribed cube of the candidate segmentation result, 10 pixels can be expanded in a direction away from the center of the maximum circumscribed cube to obtain the bounding box of the candidate segmentation result. In this example, n is 10, or it can also be a value such as 5, 8, 12, etc. The embodiment of the present application does not limit the number of pixels that are expanded outward.

[0111] S503: Determine a rotation axis of the candidate segmentation result according to multiple body diagonals of the bounding box.

[0112] In one embodiment, the process of determining the rotation axis of the candidate segmentation result based on multiple body diagonals of the bounding box may include: extending the body diagonal of the bounding box inward by n pixels to obtain the starting point and the ending point of the body diagonal; if the starting point and the ending point are both in the candidate segmentation result, then determining the body diagonal as the rotation axis of the candidate segmentation result.

[0113] After obtaining the bounding box of the candidate segmentation result, the computer device obtains the body diagonal of the bounding box, and then expands the body diagonal of the bounding box inward by n pixels. If the starting point and the ending point of the body diagonal are still in the candidate segmentation result, the body diagonal of the bounding box is determined to be the rotation axis of the candidate segmentation result.

[0114] Considering that in the above embodiment of determining the rotation axis, there may be a situation where the starting point and the ending point of the four body diagonals are not in the candidate segmentation results, in this case, the rotation axis can be determined according to the number of points on each body diagonal.

[0115] Based on this, in another embodiment, the process of determining the rotation axis of the candidate segmentation result based on multiple body diagonals of the bounding box may include: obtaining the number of points in the candidate segmentation result that are located on each body diagonal; determining the body diagonal with the largest number of points as the rotation axis, and determining the center of the body diagonal as the rotation center.

[0116] In the embodiment of the present application, the body diagonal is the body diagonal of the bounding box of the candidate segmentation result, and each body diagonal passes through the candidate segmentation result, so that a part of the points in the candidate segmentation result must be on each body diagonal. For any body diagonal, the number of points in the candidate segmentation result on the body diagonal is obtained. If the number of points on the body diagonal is the largest, the body diagonal is determined as the rotation axis, and the center of the body diagonal is determined as the rotation center. It should be understood that in this embodiment, the rotation axis is determined by the number of points on the body diagonal in order to ensure that each rotating surface has a uniform contour as much as possible.

[0117] In the volume measurement method of the target tissue provided in the embodiment of the present application, the maximum circumscribed cube of the candidate segmentation result is obtained, and then the maximum circumscribed cube is expanded outward by multiple pixels to obtain the bounding box of the candidate segmentation result, and finally the rotation axis of the candidate segmentation result is determined according to the body diagonal of the bounding box. In this method, an optional way to quickly obtain the rotation axis is provided. By obtaining the maximum circumscribed cube of the candidate segmentation result, the rotation axis of the candidate segmentation result can be determined according to the body diagonal of the maximum circumscribed cube, which provides data support for the subsequent extraction of sub-contours.

[0118] Based on the above embodiment, an embodiment is provided to illustrate the process of obtaining multiple rotation planes of the candidate segmentation results according to the rotation axis and the preset angle step.

[0119] In an exemplary embodiment, Figure 6 As shown, this embodiment includes the following steps 601 to 603:

[0120] S601, obtaining multiple vectors perpendicular to the rotation axis, and determining any one of the vectors perpendicular to the rotation axis as an initial normal vector.

[0121] The computer device obtains a vector perpendicular to the rotation axis according to the rotation axis. There are multiple vectors, and any one of the vectors can be determined as an initial normal vector to determine multiple normal vectors based on the initial normal vector.

[0122] S602, obtaining multiple candidate normal vectors starting from an initial normal vector according to the angle step.

[0123] Starting from the initial normal vector, the initial normal vector is rotated according to the preset angle step until it rotates to 180° and stops, and multiple normal vectors can be obtained, which are candidate normal vectors. Taking the preset angle step of 15° as an example, a normal vector is obtained every 15°, and a total of 11 candidate normal vectors can be obtained.

[0124] S603: Determine multiple rotation planes of the candidate segmentation results according to the rotation axis, the rotation center, the initial normal vector and each candidate normal vector.

[0125] According to the rotation axis, rotation center and normal vector, a plane can be obtained by combining the plane model expression. In this way, multiple planes can be obtained according to multiple normal vectors, that is, multiple rotation surfaces can be obtained.

[0126] Exemplarily, an initial rotation plane corresponding to an initial normal vector is determined based on a plane model expression, the rotation axis coordinates of the rotation axis, the rotation center coordinates and an initial normal vector; a candidate rotation plane corresponding to each candidate normal vector is determined based on a plane model expression, the rotation axis coordinates of the rotation axis, the rotation center coordinates and each candidate normal vector; and the initial rotation plane and each candidate rotation plane are determined as multiple rotation planes of a candidate segmentation result.

[0127] The computer device can first obtain the plane model expression, and then obtain the rotation axis coordinates and the rotation center coordinates of the rotation axis. After that, the rotation axis coordinates, the rotation center coordinates and the initial normal vector can be input into the plane model expression to obtain the initial rotation surface corresponding to the initial normal vector.

[0128] Similarly, the rotation axis coordinates, the rotation center coordinates and each candidate normal vector are input into the plane expression to obtain the candidate rotation plane corresponding to each candidate normal vector. Further, the obtained initial rotation plane and each candidate rotation plane are determined as multiple rotation planes of the candidate segmentation result.

[0129] In the volume measurement method of the target tissue provided in the embodiment of the present application, a vector perpendicular to the rotation axis is obtained, and any vector perpendicular to the rotation axis is determined as an initial normal vector, and then multiple candidate normal vectors are obtained starting from the initial normal vector according to the angle step, and finally multiple rotation planes of the candidate segmentation result are determined according to the rotation axis, the rotation center, the initial normal vector and each candidate normal vector. In this method, the initial normal vector is obtained according to the rotation axis, and multiple candidate normal vectors are obtained according to the initial normal vector. In this way, multiple planes can be obtained according to the normal vector and the plane model expression, that is, multiple rotation planes can be obtained.

[0130] Based on any of the foregoing embodiments, an embodiment is provided to illustrate a process of adjusting multiple sub-contours of a target tissue.

[0131] In an exemplary embodiment, Figure 7 As shown, the adjustment process of multiple sub-contours includes the following steps 701 to 703:

[0132] S701, displaying a contour adjustment interface.

[0133] In the embodiment of the present application, the contour adjustment interface includes multiple sub-contours of the target tissue.

[0134] When fine-tuning of the multiple sub-contours of the target tissue is required, the computer device displays a contour adjustment interface to the user, so that the user can operate and adjust the multiple sub-contours.

[0135] S702: Receive adjustment position information of a user operating on a plurality of sub-contours.

[0136] After the computer device displays the contour adjustment interface to the user, the user can review each sub-contour displayed in the contour adjustment interface, and then operate the sub-contour that needs to be adjusted based on the user's personal experience and professional knowledge. The user here may refer to a doctor or a professional familiar with tissue contours, and this embodiment of the application does not limit this.

[0137] For any sub-contour, if the sub-contour needs to be fine-tuned, the user can operate the sub-contour in the contour adjustment interface. After the user operates the sub-contour, the computer device can obtain the adjustment position information of the user's operation on the sub-contour.

[0138] S703: Adjust the multiple sub-contours according to the adjustment position information of the multiple sub-contours to generate multiple adjusted sub-contours.

[0139] After acquiring the adjustment position information of the user's operation on the multiple sub-contours, the computer device adjusts the multiple sub-contours by adjusting the position information to obtain the multiple sub-contours after fine-tuning.

[0140] In the volume measurement method of the target tissue provided in the embodiment of the present application, by displaying a contour adjustment interface, the contour adjustment interface includes multiple sub-contours of the target tissue, and then receiving the adjustment position information of the user operating the multiple sub-contours, and finally adjusting the multiple sub-contours according to the adjustment position information of the multiple sub-contours to generate the adjusted multiple sub-contours. In this method, by displaying the contour adjustment interface to the user, the user can operate the multiple sub-contours based on the contour adjustment interface to generate multiple sub-contours after fine-tuning according to the operated sub-contours. In this process, the multiple sub-contours are adjusted by effectively combining the personal experience and professional knowledge of the doctor, so that the multiple sub-contours are more accurate, the accuracy of the sub-contours is improved, and the accuracy of the volume measurement of the target tissue is improved.

[0141] Based on any of the foregoing embodiments, an embodiment is provided to illustrate the process of performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue.

[0142] In an exemplary embodiment, Figure 8 As shown, the above tissue contour determination process includes the following steps 801 to 803:

[0143] S801, obtaining pixel coordinates of a plurality of pixel points in a boundary box.

[0144] As mentioned above, the bounding box is a cube obtained by expanding the largest circumscribed cube of the candidate segmentation result by n pixels. The pixel coordinates of all pixels in the bounding box can be obtained according to the position of the bounding box.

[0145] S802, inputting the pixel coordinates of each pixel point into the classification model to obtain the classification result of each pixel point.

[0146] The classification model is a model that determines whether each pixel in the bounding box belongs to the inside or outside of the target tissue contour.

[0147] After obtaining the pixel coordinates of multiple pixel points in the bounding box, the computer device inputs the pixel coordinates of each pixel point into the classification model, and then the classification model can output the classification result of each pixel point. Before the pixel coordinates of each pixel point are input into the classification model, the pixel coordinates of each pixel point are normalized.

[0148] S803, performing real-time correction processing on the adjusted multiple sub-contours according to the classification results of each pixel point to obtain the tissue contour of the target tissue.

[0149] After obtaining the classification results of each pixel point, it can be determined whether each pixel point belongs to the target tissue contour or outside the contour, and then the target tissue can be formed according to each pixel point belonging to the contour. Furthermore, the initial tissue contour of the target tissue can be fitted according to the fine-tuned multiple sub-contours, and then the edge pixel points of the target tissue formed above are obtained, and the pixel points in the fitted initial tissue contour are corrected to obtain the corrected initial tissue contour. Naturally, the corrected initial tissue contour is the tissue contour of the target tissue.

[0150] In the volume measurement method of the target tissue provided in the embodiment of the present application, the pixel coordinates of multiple pixel points in the boundary box are obtained, and then the pixel coordinates of each pixel point are input into the classification model to obtain the classification results of each pixel point, and then the adjusted multiple sub-contours are corrected according to the classification results of each pixel point to obtain the tissue contour of the target tissue. In this method, by inputting the pixel coordinates of each pixel point into the classification model, the classification result of whether each pixel point belongs to the target tissue contour can be obtained, so that the fine-tuned multiple sub-contours can be corrected according to each pixel point belonging to the target tissue contour, so that the tissue contour of the target tissue obtained is more accurate, and the accuracy of the volume measurement of the target tissue is improved.

[0151] Based on any of the foregoing embodiments, an embodiment is provided to illustrate the process of building a classification model.

[0152] In an exemplary embodiment, Fig. 9 As shown, the construction process of the classification model includes the following steps 901 to 903:

[0153] S901, obtaining contour coordinate data of a plurality of adjusted sub-contours.

[0154] After adjusting the multiple sub-contours of the target tissue, multiple adjusted sub-contours can be obtained. The computer device can obtain the coordinates of each sub-contour according to the multiple adjusted sub-contours to obtain contour coordinate data of the multiple sub-contours.

[0155] S902, normalizing each contour coordinate data to obtain normalized contour coordinate data of a plurality of adjusted sub-contours.

[0156] Exemplarily, the depth information, height information and width information of the three-dimensional image are obtained; and the coordinate data of each contour are normalized according to the depth information, height information and width information to obtain the normalized contour coordinate data of the adjusted multiple sub-contours.

[0157] When the computer device normalizes each contour coordinate data, it can map the depth d, height h and width w of the three-dimensional image of the target tissue. Then, according to the preset mapping formula, each contour coordinate data is mapped to [-1, 1]. Among them, the mapping formula of the contour coordinates (X, Y, Z) is (2X / (w-1), 2Y / (h-1), 2Z / (d-1)).

[0158] S903, performing online training on a preset neural network according to each normalized contour coordinate data to obtain a classification model.

[0159] After obtaining the normalized contour coordinate data of multiple sub-contours, each normalized contour coordinate data can be used as a training data set to perform model training on the preset neural network to obtain a classification model. The preset neural network is a weight-initialized neural network, which can be an MLP (Multi-Layer Perceptron) network. MLP is a network composed of multiple fully connected layers and has not been pre-trained. In addition, the preset neural network can also be a convolutional neural network (CNN), a Transformer, or other neural network.

[0160] In an embodiment of the present application, positive samples and negative samples need to be constructed in the process of training the classification model, wherein each sub-contour is eroded, and then the eroded sub-contour is subtracted from the original sub-contour to obtain a positive sample; each sub-contour is expanded, and then the expanded sub-contour is subtracted from the original sub-contour to obtain a negative sample.

[0161] Based on the positive and negative samples constructed above, both the positive and negative samples can be input into the MLP network for training, and the classification model obtained by training can be verified after training. For example, in the verification process, a contour set can be fitted to obtain a fitted contour, and the fitted contour and the contour gold standard are subjected to dice calculation to obtain the dice coefficient, so as to verify the classification model obtained by training according to the dice coefficient. The contour set can be a randomly selected rotation axis in the gold standard volume data, and then 12 discrete surfaces are selected along the rotation axis, each discrete surface has a contour, and the points on each contour constitute the contour set.

[0162] In this embodiment, the classification model is obtained by online training, and the online training method is to train the model using given training samples one by one, that is, when there is a new training sample, the model parameters are updated using the new training sample. In this way, if the adjustment of the sub-contour is performed by a professional (such as a doctor), and the sub-contour is adjusted multiple times until it is accurate, the network model can continuously learn professional knowledge and doctor's experience during the online training process to make the classification ability of the classification model more accurate, thereby ensuring that the volume measurement is highly automated and adding the doctor's professional knowledge and experience to guide the measurement results, further improving the accuracy of the volume measurement.

[0163] In the volume measurement method of the target tissue provided in the embodiment of the present application, the contour coordinate data of the multiple sub-contours after fine-tuning are obtained, and then each contour coordinate data is normalized to obtain the normalized contour coordinate data of the multiple sub-contours after fine-tuning, and a preset neural network is trained according to each normalized contour coordinate data to obtain a classification model. In this method, when constructing a classification model, the contour coordinate data of the sub-contours are first normalized to obtain normalized contour coordinate data, and then the normalized contour coordinate data is used to train the neural network model to obtain a classification model, thereby improving the prediction accuracy and training efficiency of the classification model.

[0164] In addition, in an exemplary embodiment, the present application also provides an optional example of a method for measuring the volume of a target tissue, such as Fig.10 As shown, the following steps may be included:

[0165] S1001, in response to a volume measurement request for a target tissue, a three-dimensional image is input into a preset segmentation model to obtain a candidate segmentation result of the target tissue output by the segmentation model.

[0166] S1002, obtaining the largest circumscribed cube of the candidate segmentation result.

[0167] S1003, expanding the largest circumscribed cube outward by multiple pixels to obtain a bounding box of the candidate segmentation result.

[0168] S1004: Determine a rotation axis of the candidate segmentation result according to the body diagonal of the bounding box.

[0169] S1005, obtaining a vector perpendicular to the rotation axis, and determining any vector perpendicular to the rotation axis as an initial normal vector.

[0170] S1006, obtaining multiple candidate normal vectors starting from the initial normal vector according to the angle step.

[0171] S1007, determining multiple rotation planes of the candidate segmentation results according to the rotation axis, the rotation center, the initial normal vector and each candidate normal vector.

[0172] S1008, determining multiple sub-contours of the target tissue according to multiple rotation planes of the candidate segmentation results.

[0173] S1009, adjusting the multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue.

[0174] S1010, determining the volume of the target tissue according to the tissue contour of the target tissue.

[0175] The above processes of S1001 - S1010 can refer to the description of the above method embodiment, and the implementation principles and technical effects are similar, which will not be repeated here.

[0176] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0177] Based on the same inventive concept, the embodiment of the present application also provides a target tissue volume measurement device for implementing the target tissue volume measurement method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in the embodiments of the volume measurement device for one or more target tissues provided below can refer to the limitations of the volume measurement method above, and will not be repeated here.

[0178] In an exemplary embodiment, Fig.11 As shown, a target tissue volume measurement device 1 is provided, comprising: a sub-contour acquisition module 10, a first determination module 20 and a second determination module 30, wherein:

[0179] The sub-contour acquisition module 10 is used to acquire a plurality of sub-contours of the target tissue according to the three-dimensional image of the target tissue in response to a volume measurement request of the target tissue;

[0180] A first determination module 20 is used to adjust multiple sub-contours of the target tissue, and perform real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0181] The second determination module 30 is used to determine the volume of the target tissue according to the tissue contour of the target tissue.

[0182] In one embodiment, the sub-contour acquisition module 10 is further used for:

[0183] The three-dimensional image is input into a preset segmentation model to obtain candidate segmentation results of the target tissue output by the segmentation model; and multiple sub-contours of the target tissue are determined based on the candidate segmentation results.

[0184] In one embodiment, the sub-contour acquisition module 10 is further used for:

[0185] According to the candidate segmentation results, the rotation axis of the candidate segmentation results is obtained; according to the rotation axis and a preset angle step, multiple rotation planes of the candidate segmentation results are obtained; according to the multiple rotation planes of the candidate segmentation results, multiple sub-contours of the target tissue are determined.

[0186] In one embodiment, the sub-contour acquisition module 10 is further used for:

[0187] Obtain the maximum circumscribed cube of the candidate segmentation result; expand the maximum circumscribed cube outward by multiple pixels to obtain the bounding box of the candidate segmentation result; and determine the rotation axis of the candidate segmentation result according to the body diagonal of the bounding box.

[0188] In one embodiment, the sub-contour acquisition module 10 is further used for:

[0189] The number of points in the candidate segmentation results that are located on each body diagonal is obtained; the body diagonal with the largest number of points is determined as the rotation axis, and the center of the body diagonal is determined as the rotation center.

[0190] In one embodiment, the sub-contour acquisition module 10 is further used for:

[0191] Obtain multiple vectors perpendicular to the rotation axis, and determine any one of the vectors perpendicular to the rotation axis as the initial normal vector; according to the angle step, starting from the initial normal vector, obtain multiple candidate normal vectors; according to the rotation axis, the rotation center, the initial normal vector and each candidate normal vector, determine multiple rotation planes of the candidate segmentation result; the rotation center is the center of the rotation axis.

[0192] In one embodiment, the first determining module 20 is further configured to:

[0193] Displaying a contour adjustment interface; the contour adjustment interface includes multiple sub-contours of a target tissue; receiving adjustment position information of a user operating on multiple sub-contours; adjusting the multiple sub-contours according to the adjustment position information of the multiple sub-contours to generate multiple adjusted sub-contours.

[0194] In one embodiment, the first determining module 20 is further configured to:

[0195] The pixel coordinates of multiple pixel points in the bounding box are obtained; the pixel coordinates of each pixel point are input into the classification model to obtain the classification result of each pixel point; according to the classification result of each pixel point, the adjusted multiple sub-contours are corrected in real time to obtain the tissue contour of the target tissue.

[0196] In one embodiment, the volume measuring device 1 further includes:

[0197] A data acquisition module, used for acquiring contour coordinate data of the adjusted multiple sub-contours;

[0198] A third determination module is used to perform normalization processing on each contour coordinate data to obtain normalized contour coordinate data of a plurality of adjusted sub-contours;

[0199] The fourth determination module is used to perform online training on a preset neural network according to each normalized contour coordinate data to obtain a classification model.

[0200] Each module in the above volume measurement device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0201] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0202] In response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue;

[0203] Adjusting multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0204] According to the tissue contour of the target tissue, the volume of the target tissue is determined.

[0205] The implementation principles and technical effects of each step implemented by the processor in the embodiment of the present application are similar to the principles of the above-mentioned volume measurement method and will not be repeated here.

[0206] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0207] In response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue;

[0208] Adjusting multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0209] According to the tissue contour of the target tissue, the volume of the target tissue is determined.

[0210] The implementation principles and technical effects of the various steps implemented when the computer program in the embodiment of the present application is executed by the processor are similar to the principles of the above-mentioned volume measurement method and will not be repeated here.

[0211] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0212] In response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue;

[0213] Adjusting multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue;

[0214] According to the tissue contour of the target tissue, the volume of the target tissue is determined.

[0215] The implementation principles and technical effects of the various steps implemented when the computer program in the embodiment of the present application is executed by the processor are similar to the principles of the above-mentioned volume measurement method and will not be repeated here.

[0216] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0217] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0218] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0219] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for measuring the volume of a target tissue, characterized in that: The method comprises: In response to a volume measurement request for a target tissue, acquiring a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue; Adjusting the multiple sub-contours of the target tissue, and performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue; The volume of the target tissue is determined according to the tissue contour of the target tissue.

2. The method according to claim 1, characterized in that The step of acquiring a plurality of sub-contours of the target tissue according to the three-dimensional image of the target tissue comprises: Inputting the three-dimensional image into a preset segmentation model to obtain a candidate segmentation result of the target tissue output by the segmentation model; A plurality of sub-contours of the target tissue are determined according to the candidate segmentation results.

3. The method according to claim 2, characterized in that Determining a plurality of sub-contours of the target tissue according to the candidate segmentation results includes: According to the candidate segmentation result, obtaining a rotation axis of the candidate segmentation result; Acquire multiple rotation planes of the candidate segmentation result according to the rotation axis and a preset angle step; A plurality of sub-contours of the target tissue are determined according to the plurality of rotation planes of the candidate segmentation results.

4. The method according to claim 3, characterized in that The step of obtaining a rotation axis of the candidate segmentation result according to the candidate segmentation result includes: Obtain the largest circumscribed cube of the candidate segmentation result; Expand the maximum circumscribed cube outward by a plurality of pixels to obtain a bounding box of the candidate segmentation result; A rotation axis of the candidate segmentation result is determined according to a plurality of body diagonals of the bounding box.

5. The method according to claim 3, characterized in that: The step of obtaining a plurality of rotation planes of the candidate segmentation result according to the rotation axis and a preset angle step comprises: Acquire multiple vectors perpendicular to the rotation axis, and determine any one of the vectors perpendicular to the rotation axis as an initial normal vector; According to the angle step, starting from the initial normal vector, obtaining a plurality of candidate normal vectors; A plurality of rotation planes of the candidate segmentation result are determined according to the rotation axis, the rotation center, the initial normal vector and each of the candidate normal vectors.

6. The method according to any one of claims 1 to 5, characterized in that: The adjusting of the multiple sub-profiles of the target tissue comprises: Displaying a contour adjustment interface; the contour adjustment interface includes a plurality of sub-contours of the target tissue; receiving adjustment position information of the plurality of sub-contours operated by a user; The plurality of sub-contours are adjusted according to the adjustment position information of the plurality of sub-contours to generate the plurality of adjusted sub-contours.

7. The method according to any one of claims 1 to 5, characterized in that: The step of performing real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue includes: Get the pixel coordinates of multiple pixels in the bounding box; Inputting the pixel coordinates of each pixel point into the classification model to obtain the classification result of each pixel point; According to the classification results of each of the pixel points, the adjusted multiple sub-contours are corrected in real time to obtain the tissue contour of the target tissue.

8. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Acquire contour coordinate data of the adjusted multiple sub-contours; Normalizing each of the contour coordinate data to obtain normalized contour coordinate data of the adjusted multiple sub-contours; According to each of the normalized contour coordinate data, a preset neural network is trained online to obtain the classification model.

9. A device for measuring the volume of a target tissue, characterized in that: The device comprises: A sub-contour acquisition module, configured to acquire a plurality of sub-contours of the target tissue according to a three-dimensional image of the target tissue in response to a volume measurement request of the target tissue; A first determination module is used to adjust the multiple sub-contours of the target tissue, and perform real-time correction processing on the adjusted multiple sub-contours according to the online trained classification model to obtain the tissue contour of the target tissue; The second determination module is used to determine the volume of the target tissue according to the tissue contour of the target tissue.

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

Images