Image report generation method, device, system, equipment and storage medium

Abstract

The present disclosure provides an image report generation method, device, system, equipment and storage medium. The image report generation method of the present disclosure comprises: acquiring three-dimensional volume data, reconstruction geometry parameters and two-dimensional projection data of a first object from CBCT; performing identification and classification of a region of interest according to the three-dimensional volume data, the reconstruction geometry parameters and the two-dimensional projection data of the first object to obtain region of interest information of the first object; performing local section reconstruction according to the region of interest information of the first object, the reconstruction geometry parameters and the two-dimensional projection data to obtain a first report map of the first object; and generating a to-be-confirmed image report of the first object, wherein the to-be-confirmed image report of the first object contains the first report map, so as to show the to-be-confirmed image report of the first object to a first user and enable the first user to audit the to-be-confirmed image report of the first object. The present disclosure can automatically generate an image report containing a lesion region image with higher resolution.

Description

Technical Field

[0001] This disclosure relates to a method, apparatus, system, device, and storage medium for generating image reports. Background Technology

[0002] In some applications, professionals need to manually extract key parts of images, annotate and format them to generate reports, so that relevant personnel can review them and obtain key information from the images. For example, in dental imaging reports, dental radiologists need to provide patients with imaging reports based on 3D data from cone-beam computed tomography (CBCT). Doctors need to manually extract numerous slices of lesion areas using tools such as image reading software, annotate and format them, and then import them into the imaging report. Summary of the Invention

[0003] To address at least one of the aforementioned technical problems, this disclosure provides an image report generation method, apparatus, system, device, and storage medium capable of automatically generating image reports containing images of lesion areas.

[0004] The first aspect of this disclosure provides a method for generating image reports, including:

[0005] Acquire the three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object from CBCT;

[0006] Based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, the region of interest is identified and classified to obtain the region of interest information of the first object.

[0007] Based on the region of interest information of the first object, the reconstructed geometric parameters, and the two-dimensional projection data, local cross-sectional reconstruction is performed to obtain the first report image of the first object;

[0008] A pending image report for the first object is generated, which includes the accompanying image from the first report, to be displayed to the first user so that the first user can review the pending image report for the first object.

[0009] A second aspect of this disclosure provides an image report generation apparatus, comprising:

[0010] The acquisition unit is used to acquire the three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object from CBCT.

[0011] The identification and classification unit is used to identify and classify the region of interest based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, so as to obtain the region of interest information of the first object.

[0012] The local reconstruction unit is used to perform local cross-sectional reconstruction based on the region of interest information of the first object, the reconstruction geometric parameters and the two-dimensional projection data, to obtain the first report image of the first object;

[0013] A report generation unit is used to generate a report on the image to be confirmed for the first object, wherein the report on the image to be confirmed for the first object includes the image in the first report.

[0014] The image report review unit is used to provide the first user with the image report to be confirmed for the first object, so that the first user can review the image report to be confirmed for the first object.

[0015] A third aspect of this disclosure provides an electronic device comprising:

[0016] Memory, the memory storing execution instructions; and

[0017] A processor that executes the execution instructions stored in the memory, causing the processor to perform the aforementioned image report generation method.

[0018] A fourth aspect of this disclosure provides a readable storage medium storing executable instructions that, when executed by a processor, are used to implement the image report generation method described above.

[0019] The fifth aspect of this disclosure provides an image report generation system, including: a CBCT and a cloud server; wherein the cloud server includes a data storage unit, a report storage unit, and the aforementioned image report generation device; the data storage unit is used to store three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of a first object from the CBCT; the report storage unit is used to store an unconfirmed image report, a confirmed image report, and / or a visualized three-dimensional image of the first object obtained by the image report generation device.

[0020] This disclosure uses raw CBCT scan data (i.e., two-dimensional projection data), three-dimensional volume data obtained from CBCT reconstruction, and reconstructed geometric parameters to automatically identify and classify regions of interest that may belong to lesions or suspected lesions. Based on the information of the regions of interest, local cross-sectional images are generated, which can effectively improve the clarity and resolution of the images in the imaging report. This makes it easier for the first user (i.e., the doctor) to diagnose lesions by referring to the imaging report. It can free doctors from the heavy workload of writing and reviewing medical imaging reports, improve doctors' work efficiency, and also enable the second user (i.e., the patient) to intuitively and clearly understand their lesions and condition. Attached Figure Description

[0021] The accompanying drawings illustrate exemplary embodiments of the present disclosure and, together with the description thereof, serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0022] Figure 1 This is a flowchart illustrating an embodiment of the image report generation method disclosed herein.

[0023] Figure 2 This is a schematic diagram of the reconstructed geometric parameters of one embodiment of this disclosure.

[0024] Figure 3 This is a schematic diagram of an artificial intelligence model and image report generation process according to one embodiment of this disclosure.

[0025] Figure 4 This is a schematic diagram illustrating the specific implementation process of an image report generation method according to one embodiment of the present disclosure.

[0026] Figure 5 This is a schematic diagram of the MPR interface before doctor review, as described in one embodiment of this disclosure.

[0027] Figure 6 This is a schematic diagram of the MPR interface after doctor review, as one embodiment of this disclosure.

[0028] Figure 7 This is a schematic block diagram of an image report generation apparatus that employs a hardware implementation of a processing system, according to one embodiment of this disclosure.

[0029] Figure 8 This is a schematic block diagram of the structure of an image report generation system according to one embodiment of the present disclosure.

[0030] Explanation of reference numerals in the attached figures

[0031] 700 Image Report Generation Device

[0032] 702 Acquisition Unit

[0033] 704 Identification Classification Unit

[0034] 706 Local Reconstruction Unit

[0035] 708 Report Generation Unit

[0036] 710 Image Report Review Unit

[0037] 712 Rendering Units

[0038] 714 Image Report Distribution Unit

[0039] 716 model training units

[0040] 800 bus

[0041] 900 processor

[0042] 1000 memory

[0043] 1100 Various other circuits

[0044] 1200 cloud servers

[0045] 1300 CBCT or CBCT main unit

[0046] 1400 Cloud Service Adapter

[0047] 1202 Data Storage Unit

[0048] 1204 Report Storage Unit

[0049] 1208 Local Reconstruction Interface

[0050] 1206 Data Rendering Interface Detailed Implementation

[0051] The present disclosure will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the disclosure. Furthermore, it should be noted that, for ease of description, only the parts relevant to the present disclosure are shown in the accompanying drawings.

[0052] It should be noted that, where there is no conflict, the embodiments and features described in this disclosure can be combined with each other. The technical solutions of this disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0053] Unless otherwise stated, the exemplary implementations / embodiments shown are to be understood as providing exemplary features of various details that provide ways in which the technical concepts of this disclosure can be implemented in practice. Therefore, unless otherwise stated, the features of various implementations / embodiments may be additionally combined, separated, interchanged and / or rearranged without departing from the technical concepts of this disclosure.

[0054] The use of crosshairs and / or shading in the accompanying drawings is generally used to clarify the boundaries between adjacent components. Thus, unless otherwise stated, the presence or absence of crosshairs or shading does not convey or indicate any preference or requirement for the specific material, material properties, dimensions, proportions, commonalities between the illustrated components, or any other characteristics, properties, etc., of the components. Furthermore, in the accompanying drawings, the dimensions and relative dimensions of components may be exaggerated for clarity and / or descriptive purposes. When exemplary embodiments can be implemented differently, a specific process sequence may be performed in a different order than that described. For example, two consecutively described processes may be performed substantially simultaneously or in the reverse order of their description. Furthermore, the same reference numerals denote the same components.

[0055] When a component is referred to as being "on" or "above" another component, "connected to," or "joined to" another component, the component may be directly on, directly connected to, or directly joined to the other component, or there may be intermediate components. However, when a component is referred to as being "directly on" another component, "directly connected to," or "directly joined to" another component, there are no intermediate components. Therefore, the term "connection" can refer to a physical connection, an electrical connection, etc., and may or may not have intermediate components.

[0056] The terminology used herein is for the purpose of describing particular embodiments and is not intended to be limiting. As used herein, unless the context clearly indicates otherwise, the singular forms “a” and “the” are intended to include the plural forms as well. Furthermore, when the terms “comprising” and / or “including” and variations thereof are used in this specification, it indicates the presence of the stated features, integrals, steps, operations, parts, components, and / or groups thereof, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, parts, components, and / or groups thereof. It should also be noted that, as used herein, the terms “substantially,” “about,” and other similar terms are used as approximate terms rather than as terms of degree, thus explaining the inherent biases in measurements, calculated values, and / or provided values ​​that would be recognized by one of ordinary skill in the art.

[0057] In related technologies, CBCT electronic image reporting systems have limited functionality. Artificial intelligence models use two-dimensional images (e.g., panoramic or lateral views) to identify and analyze lesions to generate corresponding medical image reports. Furthermore, the image reports are limited in content, lacking high-resolution images of lesion sites or suspected lesion sites. Doctors often struggle to verify the accuracy of lesion images in the reports, frequently requiring manual modification or reconstruction, which is tedious, time-consuming, and labor-intensive. Additionally, DICOM front-end processors and cloud services are required for image interpretation. Moreover, the artificial intelligence models used for lesion identification and classification cannot self-iterate or self-learn, making the process of updating these models cumbersome and time-consuming. Therefore, this disclosure provides an image report generation method, apparatus, device, storage medium, and system.

[0058] This disclosure is applicable to various medical scenarios, such as oral and maxillofacial surgery.

[0059] The following text combines Figures 1 to 8 The specific embodiments of this disclosure will be described in detail.

[0060] Figure 1 A flowchart illustrating an image report generation method in some embodiments of this disclosure is shown. See also Figure 1 As shown, the image report generation method S10 of this disclosure may include:

[0061] Step S12: Obtain the three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object from CBCT;

[0062] The first target is the object being imaged by CBCT. For example, it can be, but is not limited to, parts of the human body that require medical imaging, such as the oral cavity, skull, ears, nose and throat, and teeth.

[0063] CBCT performs a circular digital radiograph around the first object to obtain two-dimensional projection data of the first object. The CBCT host then performs three-dimensional reconstruction on the two-dimensional projection data obtained after multiple (e.g., 180-360) digital radiographs around the first object based on the reconstruction geometry parameters to obtain three-dimensional volume data. This three-dimensional volume data is then rendered to form a visualized three-dimensional image of the first object. In step S12, the CBCT host can directly or indirectly (e.g., via cloud service adapter 1400) transmit the unrendered three-dimensional volume data, the obtained two-dimensional projection data, and the corresponding reconstruction geometry parameters to the cloud server.

[0064] Here, the two-dimensional projection data from CBCT may include, but is not limited to, 360° two-dimensional projection data of the first object. The two-dimensional projection data may be, but is not limited to, raw image data format (RAW data format). The three-dimensional volume data may be, but is not limited to, Digital Imaging and Communications in Medicine (DICOM) data or other medical image format data.

[0065] Reconstructed geometric parameters are the relevant geometric parameters of the CBCT scanning track. They can describe the geometric relationship between the source, detector, rotation axis, and imaging area in CBCT. The CBCT host can perform three-dimensional reconstruction based on the reconstructed geometric parameters.

[0066] Figure 2 The spatial geometric relationships expressed by the reconstructed geometric parameters are illustrated. For some implementations, see [link to implementation details]. Figure 2 As shown, the reconstructed geometric parameters may include, but are not limited to, the distance between the source and the detector (SID), the distance between the source and the rotation axis (SAD), the detector horizontal offset (uOffset), the detector vertical offset (vOffset), the source horizontal offset (ySrcOffset), and the source vertical offset (zSrcOffset). Figure 2 The rectangular coordinate system xyz in CBCT is the world coordinate system, and the pixel coordinate system uv is the pixel coordinate system of CBCT.

[0067] In some implementations, the CBCT host can upload the three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object to the cloud server through lossless compression, data anonymization, and other methods. This can protect user privacy without causing data distortion, while reducing data transmission bandwidth requirements, improving data utilization, and lowering hardware costs.

[0068] Step S14: Based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, perform region of interest identification and classification to obtain the region of interest information of the first object;

[0069] The region of interest (ROI) is the area corresponding to the automatically identified lesion or suspected lesion. In some implementations, ROI information may include, but is not limited to, the location of the ROI (e.g., the location of the center point of the ROI in a predetermined world coordinate system, which may be a world coordinate system with a preset point of the human body such as the head or face as its origin), dimensions (e.g., the length, width, and height parameters of the ROI in the predetermined world coordinate system), normal vector, name (e.g., the corresponding lesion name), and attributes (e.g., lesion category and lesion description). Here, the normal vector may indicate the cross-section corresponding to the ROI.

[0070] In some implementations, regions of interest (ROIs) can be identified and classified using artificial intelligence (AI) models. For example, an AI model may include a sequentially connected convolutional neural network (CNN) and a multi-label classification network, where the CNN is used to identify the ROI and the multi-label classification network is used to classify it. However, AI models can also be implemented using other types of machine learning models, not limited to CNNs.

[0071] Figure 3 The diagram illustrates the process of region of interest (ROI) identification and classification using an artificial intelligence model. The reconstructed geometric parameters, 2D projection data, and 3D volume data of a first object are processed by a convolutional neural network to obtain multiple visual features of the first object. Each visual feature may include the location, size, and normal vector of a ROI (e.g., a lesion or suspected lesion site on the first object). A multi-label classification network classifies the visual features of the first object obtained by the convolutional neural network, i.e., classifies the ROIs, generating label information for each ROI. This label information may include the ROI's identifier (e.g., lesion name, lesion number, etc.) and attributes (e.g., lesion category, lesion description, etc.).

[0072] In practical applications, artificial intelligence models can be trained using manually labeled samples, or they can be trained using previously obtained confirmed image reports and corresponding raw data (i.e., 3D volume data, 2D projection data, and reconstructed geometric parameters). For example, see... Figure 3 As shown, the AI ​​model can be iteratively updated using confirmed image reports to continuously improve its accuracy in application, thereby enhancing the accuracy of lesion area identification and classification.

[0073] See below Figure 4 As an example, in some implementations, the image report generation method of this disclosure may further include:

[0074] Step S13: Based on the confirmed image report of the first object, the 3D volume data, the reconstructed geometric parameters, and the 2D projection data, update the parameters of the artificial intelligence model to improve its accuracy. This enables the artificial intelligence model to learn and iterate automatically, continuously enhancing its stability and accuracy while achieving automatic image report generation.

[0075] In one implementation, step S13 may include:

[0076] Step a1: Use the confirmation image report of the first object, three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data to form a new sample for the artificial intelligence model, and update the sample library of the artificial intelligence model with the new sample;

[0077] Step a2: Retrain the artificial intelligence model using the updated sample library to update the parameters of the artificial intelligence model, thereby improving the accuracy of the artificial intelligence model in identifying and classifying lesions.

[0078] Here, methods such as few-shot learning or similar approaches can be used to retrain the AI ​​model based on the updated sample library. This not only updates the parameters of the AI ​​model and continuously improves its accuracy, but also reduces computational complexity and data volume, thereby lowering hardware costs.

[0079] Step S16: Based on the region of interest information, reconstructed geometric parameters and two-dimensional projection data of the first object, perform local section reconstruction to obtain the first report image of the first object;

[0080] In some implementations, the local section reconstruction process may include: reconstructing a two-dimensional section image of the required slice backup based on the location (e.g., center point location), size, normal vector, and two-dimensional projection data of the region of interest (i.e., the lesion area or suspected lesion area) of the first object, thereby obtaining the first report image of the first object. Since a local two-dimensional image is reconstructed instead of a three-dimensional image, the resource consumption during reconstruction can be greatly reduced. Furthermore, due to the lower resource consumption, the goal of quickly and efficiently reconstructing a high-definition image of a local region can be achieved.

[0081] In some implementations, the FDK filter backprojection algorithm can be used to perform local section reconstruction. That is, the FDK filter backprojection algorithm can be used to reconstruct a two-dimensional image using the reconstructed geometric parameters and local section projection data to obtain the first report image of the first object.

[0082] In some implementations, the report illustrations obtained through local section reconstruction (i.e., the first and second report illustrations herein) have a higher resolution than the slice images obtained directly from three-dimensional volume data.

[0083] Typically, a visualized 3D image is first obtained by rendering 3D volume data, and then 2D slices are obtained by directly slicing the 3D image in a certain direction or position. The resulting 2D slices have the same resolution as the 3D image. However, because the reconstruction and rendering of 3D volume data is resource-intensive and related equipment resources are limited, 3D images often have low resolution, resulting in low resolution 2D slices. Such 2D slices often affect the doctor's diagnosis due to insufficient clarity. In this disclosure, however, through local section reconstruction, high-definition section imaging of unclear lesions or suspected lesions can be achieved, improving the resolution of the section images of lesions. This makes the section images of lesions or suspected lesions have a higher resolution than 3D images. Using such section images as illustrations in imaging reports can effectively improve the clarity of lesion-related images in the reports, not only facilitating the diagnosis of lesions for doctors (i.e., the first user below) but also allowing patients (i.e., the second user below) to more clearly and intuitively view the relevant details of the lesions. Experiments have shown that local section reconstruction can improve the resolution of lesion or suspected lesion images in imaging reports from 0.25mm to 0.125mm or 0.06mm.

[0084] Step S18: Generate a confirmation image report for the first object. The confirmation image report for the first object includes a first report illustration to be displayed to the first user, so that the first user can review the confirmation image report for the first object.

[0085] In some implementations, the following text Figure 8 Taking the image report generation system shown as an example, after the cloud server generates the image report to be confirmed for the first subject, it pushes the image report to be confirmed for the first subject to the doctor (i.e., the first user) through the image report review unit. The doctor can access the image report review unit using electronic devices such as computers and mobile terminals through tools such as applications or browsers. The image report to be confirmed for the first subject is displayed to the doctor through the interface of the browser or application. In this way, the doctor can review and revise the image report in real time.

[0086] In some implementations, the generation and display of the image report to be confirmed can also be achieved using the same electronic device. This disclosure does not impose any limitations on this.

[0087] In some implementations, the region of interest information may also include information such as the category (i.e., lesion category) and attributes (e.g., lesion description, lesion name, etc.) of the region of interest. See also Figure 4As shown, the image report generation method of this disclosure may further include: step S15, generating a first text description based on the category and attributes of the region of interest, and adding the first text description to the image report to be confirmed for the first object. Thus, the automatically generated image report to be confirmed can contain both report images and corresponding text descriptions (e.g., diagnostic descriptions of lesions or suspected lesions), providing rich content without requiring manual input by the doctor, further simplifying the doctor's operation.

[0088] In some implementations, the image report review unit can be implemented as a first Web Service deployed on a cloud server. The image report review unit can provide a first webpage to a first terminal. The image report to be confirmed is displayed on the first webpage, which can only be viewed by the first user with image report review permissions. The first user can access the first webpage through a browser using the first terminal, and after logging in and authenticating, can review and revise the image report through the first webpage.

[0089] In some embodiments, the image report generation method of this disclosure may further include: providing a multiplanar reconstruction (MPR) interface, which can be used by a first user and / or a second user to display a visualized 3D image and / or a multiplanar slice image of the corresponding visualized 3D image.

[0090] In some implementations, the image report review unit may have a multiplanar reconstruction (MPR) interface, which can be used to display visualized 3D images, axial section images, sagittal section images, and coronal section images of the first object. The first user can also access a first webpage via a browser using a first terminal, and after logging in and authenticating, view the first webpage. The user can then access the MPR interface by clicking the MPR interface button on the first webpage or by other means to perform a diagnosis by viewing the visualized 3D images, axial section images, sagittal section images, and coronal section images of the first object. The user can also revise the accompanying images in the image report through selection operations on the MPR interface.

[0091] Here, axial, sagittal, and coronal slice images can be obtained directly from the visualized 3D image. This method of obtaining axial, sagittal, and coronal slice images directly from the 3D image reduces equipment wear and performance requirements, and meets the needs of doctors for real-time drag-and-drop viewing without lag.

[0092] Step S11: In response to the confirmation operation of the first user, a confirmation message corresponding to the pending image report of the first object is generated. The pending image report of the first object is marked as the confirmed image report of the first object, and then displayed to the second user. Thus, doctors can automatically generate confirmed image reports by clicking a button, further simplifying the doctor's operation. At the same time, while ensuring the accuracy and reliability of the image report content, the confirmed image report can be pushed online to the patient, effectively simplifying the doctor-patient communication process and improving its efficiency.

[0093] In some implementations, the report images in the confirming image report and / or the image report to be confirmed may record information such as lesion location, size, and normal vector. That is, the first report image in the confirming image report and / or the image report to be confirmed records the corresponding region of interest information, and / or the second report image in the confirming image report and / or the image report to be confirmed records the corresponding selected region information. In this way, patients or doctors can jump to the display interface of the report image by clicking on the report image in the confirming image report, which facilitates doctors to quickly review the image report and patients to quickly view their condition.

[0094] See Figure 4 As shown, the image report generation method S40, in addition to steps S12 to S13 mentioned above, may also include the following steps:

[0095] Step S42: Render a visualized 3D image of the first object using the 3D volume data of the first object, and display the visualized 3D image of the first object to the first user, so that the first user can review the image report to be confirmed of the first object in conjunction with the visualized 3D image of the first object.

[0096] Since 3D reconstruction is resource-intensive, and CBCT already has 3D volume data (i.e., DICOM data) for 3D reconstruction, this disclosure directly uses the DICOM data uploaded by CBCT to render a visualized 3D image of the first object. This eliminates the need to repeatedly perform 3D reconstruction on the entire first object, thereby reducing resource consumption, improving data utilization, and increasing execution efficiency.

[0097] This allows doctors to quickly locate lesions by referring to visualized 3D images, and to perform operations such as modification and confirmation on the images to be confirmed. This not only further improves diagnostic efficiency and simplifies doctors' operations, but also improves diagnostic accuracy.

[0098] It should be noted that the execution order of step S42 and the preceding steps S14 to S13 is not limited. That is, step S42 can be executed simultaneously with, before or after the preceding steps S14 to S13.

[0099] See Figure 4 As shown, the image report generation method S40 may further include the following steps:

[0100] Step S44: Based on the selected area information of the first object, the reconstructed geometric parameters and the two-dimensional projection data, perform local section reconstruction to obtain the second report map of the first object. The selected area information of the first object is generated in response to the map modification operation and / or map reconstruction operation of the first user.

[0101] Specifically, the doctor (i.e., the first user) can select the section to be reconstructed and a local area of ​​that section on the MPR interface. The selected area information of the first object can be obtained by detecting the doctor's selection operation on the MPR interface. For example, the first terminal can obtain the selected area information of the first object by detecting the doctor's selection operation on the MPR interface and transmit it to the cloud server through the image report review unit.

[0102] In some implementations, the selected region information may include, but is not limited to, the selected location, region size, and normal vector. Here, the selected location may be the position of a point or region center point selected by the doctor in the MPR interface in a predetermined world coordinate system (e.g., a world coordinate system with a preset point in the human body such as the head or face as the origin). The region size may include the length, width, height, and other parameters set by the doctor in the MPR interface or the length, width, height, and other parameters configured by default. The normal vector is the normal vector of the cross-section where the point selected by the doctor in the MPR interface is located.

[0103] Step S46: Adjust the image report to be confirmed for the first object according to the image in the second report, so that the image report to be confirmed for the first object includes the image in the second report;

[0104] The pending image report for the first object may include, but is not limited to:

[0105] 1) Replace the first report image selected by the first user in the image report to be confirmed for the first object with the second report image to correct the error in the first report image;

[0106] 2) Add the second report image to the first subject's unconfirmed image report to make the first subject's unconfirmed image report more comprehensive and clear, so that the lesion can be clearly shown through the second report image and the first report image.

[0107] Therefore, when physicians review and revise imaging reports, if some images are unclear or inaccurate in location, affecting the diagnosis, or if images of certain lesions are omitted from the imaging report, local cross-sectional reconstruction can be performed on one or more areas through steps S44 to S46 to obtain local cross-sectional images with higher resolution and clarity. These images can then be used to replace or add to the imaging report to be confirmed if there may be errors, thereby outputting a higher quality imaging report.

[0108] Figure 5 The diagram illustrates the MPR (Mastering Performance Reconstruction) interface before doctor review. Doctors can access and log in to the first webpage via a browser. Through the MPR interface on this webpage, they can view visualized 3D images, axial section images, sagittal section images, and coronal section images of the teeth. If the doctor finds the lesion area unclear in the image report, they can select a suitable lesion area using the crosshair in the upper left corner of the MPR interface. After the doctor's selection, the cloud server automatically performs local section reconstruction of the selected lesion area. After the doctor's review and revision, the image report image more accurately reflects the lesion, and the MPR interface also more clearly displays the cross-section of the lesion. Figure 6 The MPR interface after doctor review is shown, compared to... Figure 5 , Figure 6 The MPR interface allows for a more accurate and clear display of the tooth's condition via cross-sections. Therefore, doctors can not only view visualized 3D images, axial section images, sagittal section images, and coronal section images of the primary object within the MPR interface, but also modify report illustrations using simple operations such as manipulating the crosshairs to reselect the lesion area or section.

[0109] See Figure 4 As shown, the image report generation method S40 may further include the following steps:

[0110] Step S48: Adjust the image report to be confirmed of the first object according to the second text description of the first object, so that the image report to be confirmed of the first object contains the second text description; wherein, the second text description of the first object is generated in response to the text editing operation of the first user.

[0111] Specifically, the doctor (i.e., the first user) can enter or edit text descriptions of the images in the image report review interface. A second text description of the first object can be obtained by detecting the doctor's text editing operations on the image report review interface. For example, the first terminal can obtain the second text description of the first object by detecting the doctor's text editing operations on the image report review interface and transmit it to the cloud server through the image report review unit.

[0112] In some implementations, the second textual description may include, but is not limited to, information such as the diagnosis of the lesion and the type of lesion.

[0113] In some implementations, adjusting the image report to be confirmed for the first object based on the second textual description of the first object may include, but is not limited to:

[0114] 1) Replace the first text description selected by the first user in the image report to be confirmed of the first object with the second text description to correct the error in the first text description;

[0115] 2) Add the second text description to the image report to be confirmed for the first object to make the image report to be confirmed for the first object more comprehensive and clear, and to make it easier to explain the condition of the lesion more clearly through the second text description and the first text description.

[0116] Therefore, when physicians review and revise imaging reports, if the text descriptions of some images are unclear or inaccurate, or if the text descriptions of some images are missing from the imaging report, the text descriptions in the imaging report can be modified through step S48, thereby outputting a higher quality imaging report.

[0117] See Figure 4 As shown, the image report generation method S40 may further include the following steps:

[0118] Step S41: In response to the second user's viewing operation, display a confirmation image report and / or a visualized 3D image of the first object to the second user.

[0119] In some implementations, the following text Figure 8 Taking the image report generation system shown as an example, after the cloud server generates a confirmation image report for the first object, it provides the confirmation image report for the first object to the patient (i.e., the second user) through the image report distribution unit. The patient can access the image report distribution unit using electronic devices such as computers and mobile terminals through tools such as applications or browsers. The patient performs a viewing operation in the interface of the application or browser, and the application or browser sends a request to the image distribution system of the cloud server. The image distribution system of the cloud server can then push the confirmation image report for the first object to the interface of the browser or application, and the patient can view the confirmation image report for the first object on the interface of the browser or application. Thus, it is convenient and fast for patients to view image reports in real time on various electronic devices. In addition, the viewing and display of confirmation image reports can also be achieved through other methods, which are not limited in this disclosure.

[0120] In some implementations, the image report distribution unit can be implemented as a second web service deployed on a cloud server. The image report distribution unit can provide a second webpage to a second terminal, confirming that the image report is displayed to a first user with viewing permissions through the second webpage. For example, the second user can access the second webpage using a browser or application on the second terminal. After logging in and authenticating, the user can view the second webpage, which supports operations such as film printing and image report browsing. The second terminal can be, but is not limited to, portable electronic devices, mobile devices, or other types of terminals.

[0121] In some implementations, the image report distribution unit may also have an MPR interface, which is used to display visualized 3D images, axial section images, sagittal section images, and coronal section images of the first object. After logging into the second webpage using a second terminal, the second user can jump to the MPR interface by clicking the MPR interface button on the second webpage or by other means to view the visualized 3D images, axial section images, sagittal section images, and coronal section images of the first object.

[0122] Currently, DICOM data format remains the standard medical imaging data format. However, this format is complex, requiring specialized software for parsing and viewing, making it less universally applicable to patients. Furthermore, most cloud service systems primarily function as backups of hospital data and physician services, and only support DICOM, making it cumbersome for ordinary patients to browse and view medical imaging reports. In this disclosure, the generation of imaging reports and the rendering of visualized 3D images are decoupled from the viewing process. This allows various mobile devices to view imaging reports and related images through applications or browsers. The visualized 3D images are rendered on the server side, allowing even mobile devices with limited GPU performance or portable electronic devices (e.g., laptops) to smoothly preview lesion-related images. Additionally, with the high transmission speed and low latency of 5G technology, the browsing experience on mobile devices will be indistinguishable from that on fixed electronic devices. Therefore, this disclosure, by providing imaging reports and visualized 3D images to patients through an imaging report distribution unit, greatly facilitates patients' access to medical imaging reports.

[0123] Figure 7 This is a schematic block diagram of an image report generation apparatus that employs a hardware implementation of a processing system, according to one embodiment of this disclosure.

[0124] The apparatus may include corresponding modules that perform one or more steps in the flowchart above. Therefore, each or more steps in the flowchart above can be performed by a corresponding module, and the apparatus may include one or more of these modules. A module may be one or more hardware modules specifically configured to perform a corresponding step, or implemented by a processor configured to perform a corresponding step, or stored in a computer-readable medium for implementation by a processor, or implemented through some combination thereof.

[0125] This hardware architecture can be implemented using a bus architecture. The bus architecture can include any number of interconnect buses and bridges, depending on the specific application and overall design constraints of the hardware. Bus 800 connects various circuits, including one or more processors 900, memory 1000, and / or hardware modules. Bus 800 can also connect various other circuits 1100, such as peripherals, voltage regulators, power management circuits, external antennas, etc.

[0126] Bus 800 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, this diagram uses only one connection line, but this does not imply that there is only one bus or one type of bus.

[0127] Any process or method description in the flowcharts or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain. The processor performs the various methods and processes described above. For example, the method embodiments of this disclosure may be implemented as software programs tangibly contained in a machine-readable medium, such as memory. In some embodiments, part or all of the software program may be loaded and / or installed via memory and / or a communication interface. When the software program is loaded into memory and executed by the processor, one or more steps of the methods described above may be performed. Alternatively, in other embodiments, the processor may be configured to perform one of the methods described above by any other suitable means (e.g., by means of firmware).

[0128] The logic and / or steps represented in the flowchart or otherwise described herein may be specifically implemented in any readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0129] For the purposes of this specification, a "readable storage medium" can be any means capable of containing, storing, communicating, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable read-only memory (CDROM). Furthermore, a readable storage medium can even be paper or other suitable media on which a program can be printed, since a program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in memory.

[0130] It should be understood that various parts of this disclosure can be implemented in hardware, software, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0131] Those skilled in the art will understand that all or part of the steps of the methods described above can be implemented by a program instructing related hardware. The program can be stored in a readable storage medium, and when executed, the program includes one or a combination of the steps of the method implementation.

[0132] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a single processing module, or each unit can exist physically separately, or two or more units can be integrated into a single module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a readable storage medium. The storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0133] like Figure 7 As shown, an image report generation apparatus 700 according to one embodiment of the present disclosure may include:

[0134] The acquisition unit 702 is used to acquire the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object from CBCT.

[0135] The identification and classification unit 704 is used to perform identification and classification of the region of interest based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, so as to obtain the region of interest information of the first object.

[0136] The local reconstruction unit 706 is used to perform local section reconstruction based on the region of interest information of the first object, the reconstruction geometric parameters and the two-dimensional projection data to obtain the first report image of the first object;

[0137] The report generation unit 708 is used to generate a report on the image to be confirmed for the first object, and the report on the image to be confirmed for the first object includes the image in the first report.

[0138] The image report review unit 710 is used to provide the first user with the image report to be confirmed for the first object, so that the first user can review the image report to be confirmed for the first object.

[0139] In some implementations, the image report review unit 710 can also be used to mark the image report to be confirmed of the first object as the confirmed image report of the first object in response to the confirmation operation of the first user, so as to display the confirmed image report of the first object to the second user.

[0140] In some embodiments, the image report generation apparatus 700 may further include: a rendering unit 712, used to render a visualized three-dimensional image of the first object using the three-dimensional volume data of the first object; and an image report review unit 710, which may also be used to provide a visualized three-dimensional image of the first object to a first user, so that the first user can review the image report to be confirmed of the first object in conjunction with the visualized three-dimensional image of the first object.

[0141] In some embodiments, the local reconstruction unit 706 can also be used to perform local cross-sectional reconstruction based on the selected area information of the first object, reconstruction geometric parameters and two-dimensional projection data to obtain a second report illustration of the first object; the report generation unit 708 can also be used to adjust the pending image report of the first object based on the second report illustration, so that the pending image report of the first object includes the second report illustration; wherein, the selected area information of the first object is generated in response to the illustration modification operation and / or illustration reconstruction operation of the first user.

[0142] In some implementations, the illustration modification and / or illustration reconstruction operations are performed on the display interface of the visualized 3D image of the first object.

[0143] In some implementations, the region of interest information includes the category and attributes of the region of interest; the report generation unit 708 can also be used to generate a first text description based on the category and attributes of the region of interest, and add the first text description to the pending image report of the first object.

[0144] In some implementations, the report generation unit 708 can also be used to adjust the image report to be confirmed of the first object according to the second text description of the first object, so that the image report to be confirmed of the first object includes the second text description; wherein the second text description of the first object is generated in response to the text editing operation of the first user.

[0145] In some implementations, the image report generation system may further include: an image report distribution unit 714, which, in response to a second user's viewing operation, provides the second user with a confirmed image report and / or a visualized 3D image of the first object.

[0146] In some implementations, the image report distribution unit 714 can also be used to provide an MPR interface, which can be used to display a visualized 3D image and / or a multi-planar slice image of the corresponding visualized 3D image to a first user and / or a second user.

[0147] In some embodiments, the image report generation system may further include a model training unit 716. Specifically, the identification and classification unit 704 is used to identify and classify regions of interest using an artificial intelligence model; the model training unit 716 can be used to update the parameters of the artificial intelligence model based on the confirmed image report of the first object, three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data, in order to improve the accuracy of the artificial intelligence model.

[0148] In some implementations, the artificial intelligence model includes a sequentially connected convolutional neural network and a multi-label classification network, wherein the convolutional neural network is used to identify the region of interest, and the multi-label classification network is used to classify the region of interest.

[0149] In some implementations, the local reconstruction unit 706 can be specifically used to perform local section reconstruction using the FDK filter back projection algorithm.

[0150] In some implementations, the report generation unit 708 is further configured to record region of interest information in the first report image of the image to be confirmed; and / or, record selected region information in the second report image of the image to be confirmed.

[0151] In some implementations, the image report review unit 710 can also be used to record region of interest information in the first report illustration of the confirmed image report; and / or, to record selected region information in the second report illustration of the confirmed image report.

[0152] In practical applications, the image report generation device 700 and its various units can be implemented through software, hardware, or a combination of both.

[0153] like Figure 8 As shown, an image report generation system according to one embodiment of the present disclosure may include: a cloud server 1200, a CBCT or a CBCT host 1300;

[0154] The cloud server 1200 may include a data storage unit 1202, a report storage unit 1204, and the image report generation device 700 described above;

[0155] Data storage unit 1202 can be used to store three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object from CBCT or CBCT host 1300;

[0156] The report storage unit 1204 can be used to store the pending image report, the confirmed image report, and / or the visualized three-dimensional image of the first object obtained by the image report generation device 700.

[0157] In some implementations, the cloud server 1200 may further include: a local reconstruction interface 1208 and a data rendering interface 1206; wherein, the identification and classification unit 704, the image report review unit 710, and the report storage unit 1204 can respectively call the local reconstruction unit 706 through the local reconstruction interface 1208. The image report review unit 710, the image report distribution unit 714, and / or the report storage unit 1204 can respectively call the rendering unit 712 through the data rendering interface 1206.

[0158] In some embodiments, the image report generation system may further include a cloud service adapter 1400, which can be used to transmit three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of a first object from a CBCT or CBCT host 1300 to a cloud server 1200. That is, the cloud service adapter 1400 is used for communication between the CBCT or CBCT host 1300 and the cloud server 1200, transmitting data from the CBCT or CBCT host 1300 to the cloud server 1200, and supports the transmission of DICOM data, RAM data, etc. In specific applications, the cloud service adapter 1200 can communicate with conventional imaging equipment (e.g., a CBCT or CBCT host 1300) using the DICOM protocol, or it can use a special data interface to transmit other data.

[0159] In some implementations, the image report review unit 710 may be a first Web service used to enable interaction between the image report generation device 700 and the first user (i.e., the doctor). Details regarding the first Web service can be found in the preceding descriptions and will not be repeated here.

[0160] In some implementations, the image report distribution unit 714 may be a second Web service for enabling interaction between the image report generation device 700 and a second user (i.e., the patient). Details regarding the second Web service can be found in the preceding descriptions and will not be repeated here.

[0161] In some implementations, the data storage unit 1202 and the report storage unit 1204 can be different storage spaces in the same cloud data storage device; or, the data storage unit 1202 and the report storage unit 1204 can be different cloud data storage devices.

[0162] In some implementations, the identification and classification unit 704 may be, but is not limited to, an artificial intelligence model deployed on the cloud server 1200. Details regarding the artificial intelligence model can be found in the preceding descriptions and will not be repeated here.

[0163] The image report generation system disclosed herein, when in use, after CBCT completes the imaging scan, uploads patient-related data (including DICOM data generated after 3D reconstruction, 2D projection data, and reconstructed geometric parameters) to the data storage unit via a cloud service adapter after lossless compression and data anonymization. An identification and classification unit based on an artificial intelligence model scans and analyzes this data, intelligently screening for lesions or suspected lesions and generating a pending image report. This pending image report is then stored in the report storage unit. The image report review unit pushes the pending image report to the doctor for review and revision. After the doctor's revision and confirmation, a confirmed image report is generated and output to the patient through the image report distribution unit, allowing the patient to view an accurate medical image report. Simultaneously, the confirmed image report can be fed into the identification and classification unit via, for example, a cloud service feedback interface, for self-learning and iteration to continuously strengthen the stability and accuracy of the artificial intelligence model, thereby improving the accuracy of lesion identification by the identification and classification unit.

[0164] This disclosed image report generation system combines AI-powered automatic medical image report generation technology with cloud service technology, innovatively adding a variety of user-friendly and doctor-friendly functions. It not only allows AI neural networks to achieve self-learning and self-iteration on this system, but also decouples the display method of medical images from hardware and software, and adds functions that can better display image results and data. This greatly facilitates patients and improves doctors' work efficiency, freeing doctors from the heavy workload of writing and reviewing medical image reports, and enabling patients to intuitively and clearly understand their lesions and conditions.

[0165] This disclosure also provides an electronic device, including: a memory storing execution instructions; and a processor or other hardware module executing the execution instructions stored in the memory, causing the processor or other hardware module to perform the above-described image report generation method.

[0166] This disclosure also provides a readable storage medium storing executable instructions, which, when executed by a processor, are used to implement the above-described image report generation method.

[0167] In the description of this specification, the references to terms such as "one embodiment / mode," "some embodiments / modes," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment / mode or example is included in at least one embodiment / mode or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment / mode or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments / modes or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments / modes or examples described in this specification, as well as the features of different embodiments / modes or examples.

[0168] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0169] Those skilled in the art should understand that the above embodiments are merely for illustrating the present disclosure and are not intended to limit the scope of the disclosure. Those skilled in the art can make other changes or modifications based on the above disclosure, and these changes or modifications still fall within the scope of the present disclosure.

Claims

1. A method for generating image reports, characterized in that, include: The three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object are obtained from CBCT. The two-dimensional projection data is the original data of the CBCT scan, and the three-dimensional volume data is obtained by CBCT through three-dimensional reconstruction of the two-dimensional projection data obtained after multiple digital imaging around the first object according to the reconstructed geometric parameters. Based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, the region of interest is identified and classified to obtain the region of interest information of the first object. Based on the region of interest information of the first object, the reconstructed geometric parameters, and the two-dimensional projection data, local section reconstruction is performed to obtain a first report image of the first object. This includes: reconstructing a two-dimensional section image of the required slice backup based on the position, size, normal vector, and two-dimensional projection data of the region of interest of the first object to obtain a first report image of the first object; the local section reconstruction results in a local two-dimensional image. A pending image report for the first object is generated, which includes the accompanying image from the first report, to be displayed to the first user so that the first user can review the pending image report for the first object.

2. The image report generation method according to claim 1, characterized in that, Also includes; In response to the confirmation operation of the first user, a confirmation message corresponding to the pending image report of the first object is generated, and the pending image report of the first object is marked as the confirmed image report of the first object, so as to display the confirmed image report of the first object to the second user.

3. The image report generation method according to claim 1, characterized in that, It also includes: using the three-dimensional volume data of the first object to render a visualized three-dimensional image of the first object, so as to display the visualized three-dimensional image of the first object to the first user, so that the first user can review the image report to be confirmed of the first object in conjunction with the visualized three-dimensional image of the first object.

4. The image report generation method according to claim 1, characterized in that, Also includes: Based on the selected region information of the first object, the reconstruction geometric parameters, and the two-dimensional projection data, local cross-sectional reconstruction is performed to obtain a second report image of the first object; Adjust the pending image report of the first object according to the accompanying image of the second report, so that the pending image report of the first object includes the accompanying image of the second report; The selected area information of the first object is generated in response to the first user's image modification operation and / or image reconstruction operation.

5. The image report generation method according to claim 4, characterized in that, The image modification and / or image reconstruction operations are performed on the display interface of the visualized 3D image of the first object.

6. The image report generation method according to claim 1, characterized in that, The region of interest information includes the category and attributes of the region of interest; The step of generating the image report to be confirmed for the first object includes: generating a first text description based on the category and attributes of the region of interest, and adding the first text description to the image report to be confirmed for the first object.

7. The image report generation method according to claim 1 or 6, characterized in that, Also includes: Adjust the image report to be confirmed of the first object according to the second text description of the first object, so that the image report to be confirmed of the first object includes the second text description. The second text description of the first object is generated in response to the text editing operation of the first user.

8. The image report generation method according to claim 3, characterized in that, Also includes: In response to a viewing action by a second user, a confirmation image report and / or a visualized 3D image of the first object are displayed to the second user.

9. The image report generation method according to claim 2, characterized in that, The identification and classification of the region of interest are achieved through an artificial intelligence model; The method further includes updating the parameters of the artificial intelligence model based on the confirmed image report of the first object, three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data, so as to improve the accuracy of the artificial intelligence model.

10. The image report generation method according to claim 9, characterized in that, The artificial intelligence model includes a convolutional neural network and a multi-label classification network connected in sequence. The convolutional neural network is used to identify the region of interest, and the multi-label classification network is used to classify the region of interest.

11. The image report generation method according to claim 1 or 4, characterized in that, The local section reconstruction was performed using the FDK filtering back projection algorithm.

12. The image report generation method according to claim 4, characterized in that, The region of interest information is recorded in the first report image of the confirmed image report and / or the first report image of the image to be confirmed; and / or, the selected region information is recorded in the second report image of the confirmed image report and / or the second report image of the image to be confirmed.

13. The image report generation method according to claim 3, characterized in that, Also includes: A multi-planar reconstruction (MPR) interface is provided, which is used by a first user and / or a second user to display the visualized 3D image and / or multi-planar slice images corresponding to the visualized 3D image.

14. An image report generation device, characterized in that, include: The acquisition unit is used to acquire the three-dimensional volume data, reconstructed geometric parameters, and two-dimensional projection data of the first object from CBCT. The two-dimensional projection data is the original data from the CBCT scan, and the three-dimensional volume data is obtained by the CBCT through three-dimensional reconstruction of the two-dimensional projection data obtained after multiple digital imaging around the first object based on the reconstruction geometric parameters. The identification and classification unit is used to identify and classify the region of interest based on the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object, so as to obtain the region of interest information of the first object. The local reconstruction unit is used to perform local section reconstruction based on the region of interest information of the first object, the reconstruction geometric parameters, and the two-dimensional projection data to obtain a first report image of the first object. This includes: reconstructing a two-dimensional section image of the required slice backup based on the position, size, normal vector, and two-dimensional projection data of the region of interest of the first object to obtain a first report image of the first object; the local section reconstruction results in a local two-dimensional image. A report generation unit is used to generate a report on the image to be confirmed for the first object, wherein the report on the image to be confirmed for the first object includes the image in the first report. The image report review unit is used to provide the first user with the image report to be confirmed for the first object, so that the first user can review the image report to be confirmed for the first object.

15. The image report generation apparatus according to claim 14, characterized in that, The image report review unit is further configured to, in response to the confirmation operation of the first user, generate a confirmation message corresponding to the image report to be confirmed for the first object, mark the image report to be confirmed for the first object as the confirmed image report for the first object, so as to display the confirmed image report for the first object to the second user.

16. The image report generation apparatus according to claim 14, characterized in that, Also includes: A rendering unit is used to render a visualized three-dimensional image of the first object using the three-dimensional volume data of the first object. The image report review unit is also used to provide a visualized 3D image of the first object to the first user, so that the first user can review the image report to be confirmed of the first object in conjunction with the visualized 3D image of the first object.

17. The image report generation apparatus according to claim 14, characterized in that, The local reconstruction unit is further configured to perform local section reconstruction based on the selected area information of the first object, the reconstruction geometric parameters and the two-dimensional projection data, to obtain a second report illustration of the first object; The report generation unit is further configured to adjust the image report to be confirmed of the first object according to the image in the second report, so that the image report to be confirmed of the first object includes the image in the second report; The selected area information of the first object is generated in response to the first user's image modification operation and / or image reconstruction operation.

18. The image report generation apparatus according to claim 17, characterized in that, The image modification and / or image reconstruction operations are performed on the display interface of the visualized 3D image of the first object.

19. The image report generation apparatus according to claim 16, characterized in that, The region of interest information includes the category and attributes of the region of interest; The report generation unit is further configured to generate a first text description based on the category and attributes of the region of interest, and add the first text description to the pending image report of the first object.

20. The image report generation apparatus according to claim 16 or 19, characterized in that, The report generation unit is further configured to adjust the image report to be confirmed of the first object according to the second text description of the first object, so that the image report to be confirmed of the first object includes the second text description; The second text description of the first object is generated in response to the text editing operation of the first user.

21. The image report generation apparatus according to claim 16, characterized in that, Also includes: The image report distribution unit is used to provide the second user with a confirmed image report and / or a visualized 3D image of the first object in response to the second user's viewing operation.

22. The image report generation apparatus according to claim 15, characterized in that, The identification and classification unit is specifically used to identify and classify the region of interest through an artificial intelligence model. The model training unit is used to update the parameters of the artificial intelligence model based on the confirmed image report of the first object, three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data, so as to improve the accuracy of the artificial intelligence model.

23. The image report generation apparatus according to claim 22, characterized in that, The artificial intelligence model includes a convolutional neural network and a multi-label classification network connected in sequence. The convolutional neural network is used to identify the region of interest, and the multi-label classification network is used to classify the region of interest.

24. The image report generation apparatus according to claim 14 or 17, characterized in that, The local reconstruction unit is specifically used to perform the local section reconstruction using the FDK filter back projection algorithm.

25. The image report generation apparatus according to claim 17, characterized in that, The report generation unit is further configured to record the region of interest information in the first report image of the image to be confirmed; and / or, record the selected region information in the second report image of the image to be confirmed. The image report review unit is also used to record the region of interest information in the first report illustration of the confirmed image report; and / or to record the selected region information in the second report illustration of the confirmed image report.

26. The image report generation apparatus according to claim 21, characterized in that, The image report distribution unit is also used to provide a multi-planar reconstruction (MPR) interface, which is used to display the visualized 3D image and / or the multi-planar slice image corresponding to the visualized 3D image to a first user and / or a second user.

27. An electronic device, characterized in that, include: The memory stores execution instructions; as well as A processor that executes execution instructions stored in the memory, causing the processor to perform the image report generation method according to any one of claims 1 to 13.

28. A readable storage medium, characterized in that, The readable storage medium stores execution instructions, which, when executed by a processor, are used to implement the image report generation method according to any one of claims 1 to 13.

29. An image report generation system, characterized in that, include: CBCT and cloud server; wherein the cloud server includes a data storage unit, a report storage unit, and an image report generation device as described in any one of claims 14 to 26; The data storage unit is used to store the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object from the CBCT; The report storage unit is used to store the pending image report, confirmed image report, and / or visualized 3D image of the first object obtained by the image report generation device.

30. The image report generation system according to claim 29, characterized in that, The cloud server also includes: a partial reconstruction interface and a data rendering interface; The identification and classification unit, the image report review unit, and the report storage unit respectively call the local reconstruction unit through the local reconstruction interface; The image report review unit, the image report distribution unit, and / or the report storage unit respectively call the rendering unit through the data rendering interface.

31. The image report generation system according to claim 29, characterized in that, The image report generation system further includes a cloud service adapter, which is used to transmit the three-dimensional volume data, reconstructed geometric parameters and two-dimensional projection data of the first object from the CBCT to the cloud server.

Images