Intelligent imaging scanning method, system and device, storage medium and program product

Abstract

The invention relates to the technical field of imaging scanning, in particular to an intelligent imaging scanning method, system and device, a storage medium and a program product, and the method comprises the steps: obtaining a scanning result of a to-be-scanned individual under a current scanning protocol; determining a disease recognition result based on the scanning result and a pre-established image disease recognition model; under the condition that the disease recognition result is that the disease exists, retrieving a target scanning protocol corresponding to the disease type in the disease recognition result from a pre-established scanning protocol database; judging whether supplementary scanning is needed or not based on the current scanning protocol and the target scanning protocol; when it is determined that supplementary scanning needs to be carried out, generating a supplementary scanning protocol; and continuing scanning based on the supplementary scanning protocol until the scanning is completed. According to the method, the scanning protocol can be intelligently adjusted without excessively depending on professional knowledge and experience of scanning technicians, and the conditions of scanning omission and supplementary scanning are avoided.

Description

Technical Field

[0001] The present invention relates to the field of imaging scanning technology, and in particular to an intelligent imaging scanning method, system, equipment, storage medium and program product. Background Art

[0002] In medical imaging examinations, computed tomography (CT) and magnetic resonance imaging (MRI) are generally used to image human bones or soft tissues.

[0003] Currently, CT and MRI examinations have become a common clinical method for screening and diagnosing certain diseases, and the imaging results directly influence the diagnostic analysis results of medical staff. During the scanning process, scanning technicians need to adjust the scanning information in real time based on the scan image. Adjusting the scanning information requires a high level of professional knowledge and experience. For inexperienced scanning technicians, it is difficult to accurately select and adjust the optimal scanning information for different patients, different disease states, and different scanning areas, which can easily lead to missed scans and the need for additional scans. In serious cases, the lack of scan results can affect the diagnosis and delay the patient's optimal treatment time. Summary of the Invention

[0004] In view of this, the present invention provides an intelligent imaging scanning method, system, device, storage medium and program product to solve the problem in the prior art that scanning is overly dependent on the professional knowledge and experience of scanning technicians during scanning, which easily leads to missing scanning results and affects the diagnosis results.

[0005] In a first aspect, the present invention provides an intelligent imaging scanning method, the method comprising:

[0006] Obtain the scan results of the individual to be scanned under the current scanning protocol;

[0007] Determine disease identification results based on the scan results and a pre-established imaging disease identification model;

[0008] When the disease identification result indicates that the disease exists, a target scanning protocol corresponding to the disease type in the disease identification result is retrieved from a pre-established scanning protocol database; the scanning protocol database includes scanning protocols corresponding to different disease types one by one;

[0009] Based on the current scanning protocol and the target scanning protocol, determine whether a supplementary scan is needed;

[0010] If it is determined that a supplementary scan is required, generating a supplementary scan protocol;

[0011] Scanning continues based on the supplemental scan protocol until the scan is complete.

[0012] The intelligent imaging scanning method provided by the present invention eliminates the need for over-reliance on the expertise and experience of scanning technicians during scanning. Instead, it intelligently and adaptively adjusts the scanning protocol to suit different patients, disease states, and scanning sites, ensuring that each patient receives the most comprehensive and accurate scan results, eliminating missed scans and the need for additional scans. Furthermore, the disease recognition results determined by the imaging disease recognition model can assist medical staff in diagnosis, preventing missed or misjudgment.

[0013] In an optional embodiment, the current scanning protocol includes multiple scanning sequences, and the disease identification result is determined based on the scanning results and a pre-established imaging disease identification model, including:

[0014] The scanning results corresponding to each scanning sequence are sequentially input into the imaging disease recognition model to obtain a one-to-one corresponding prediction result for each scanning sequence;

[0015] Based on all the prediction results, the disease identification result is determined.

[0016] In this embodiment, the final disease identification result is comprehensively determined based on the prediction results of all scanning sequences, which can effectively increase the prediction accuracy, thereby improving the accuracy and comprehensiveness of the scanning results.

[0017] In an optional embodiment, the scanning protocol database further includes scanning protocols corresponding to different scanning information, and the current scanning protocol is determined by the following steps:

[0018] Obtain a checklist of individuals to be scanned;

[0019] Identify scanned information in the checklist;

[0020] A current scan protocol for the individual to be scanned is determined based on the scan information and a scan protocol database.

[0021] In this embodiment, the corresponding scanning protocol can be automatically determined based on the patient's examination list, and the scan can be automatically completed, which reduces the dependence on the professional knowledge and experience of the scanning technician and improves the efficiency and accuracy of the scan.

[0022] In an optional embodiment, the current scanning protocol includes a scout image scanning sequence, and the method further includes:

[0023] Performing a scout image scan based on a scout image scan sequence to obtain a scout image scan result;

[0024] Input the scouting image scan results into the imaging disease recognition model to determine whether there are possible abnormal areas;

[0025] When it is determined that there is a possible abnormal area, the scanning starting point and scanning range of other scanning sequences in the current scanning protocol are updated;

[0026] Other scanning sequences are scanned based on the updated scanning starting point and scanning range to obtain corresponding scanning results.

[0027] In this embodiment, a positioning scan is performed before the actual scan to determine the accurate scan starting point and scanning range, which can initially ensure the comprehensiveness of the scanning range. On this basis, a supplementary scan can further improve the completeness of the scan, achieving a double guarantee of the comprehensiveness of the scanning range and preventing missed scans and the need for supplementary scans.

[0028] In an optional embodiment, the imaging disease recognition model is established by the following steps:

[0029] Acquire a data set, which includes standard images and lesion images;

[0030] Constructing an initial network structure, where the initial network structure is a multimodal model that combines a convolutional neural network, a graph convolutional network, and a multi-task learning model;

[0031] The initial network structure is trained based on the data set to obtain an imaging disease recognition model.

[0032] In this embodiment, the use of a multimodal model can effectively improve the recognition accuracy of the final imaging disease recognition model, thereby effectively improving the accuracy and comprehensiveness of the scanning range.

[0033] In an optional embodiment, after completing the scan, the method further includes:

[0034] The disease identification results corresponding to the individual to be scanned and the relevant scanning protocols are associated with the scanning protocol database.

[0035] In this embodiment, the disease identification results corresponding to the individual to be scanned and the relevant scanning protocols are associated with the scanning protocol database, which can optimize the subsequent diagnosis and treatment process and make the personalized medical plan more accurate and efficient.

[0036] In a second aspect, the present invention provides an intelligent imaging scanning system, the system comprising:

[0037] An acquisition module is used to obtain the scan result of the individual to be scanned under the current scanning protocol;

[0038] An imaging diagnosis auxiliary module is used to determine the disease identification results based on the scan results and the pre-established imaging disease identification model;

[0039] A retrieval module is used to retrieve a target scanning protocol corresponding to the disease type in the disease identification result from a pre-established scanning protocol database when the disease identification result indicates the presence of the disease; the scanning protocol database includes scanning protocols corresponding to different disease types;

[0040] A judgment module, used to judge whether a supplementary scan is required based on the current scan protocol and the target scan protocol;

[0041] A generating module, configured to generate a supplementary scanning protocol when it is determined that a supplementary scanning is required;

[0042] The scanning module is used to continue scanning based on the supplementary scanning protocol until the scanning is completed.

[0043] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to thereby execute the intelligent imaging scanning method of the first aspect or any corresponding embodiment thereof.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions for enabling a computer to execute the intelligent imaging scanning method of the first aspect or any corresponding embodiment thereof.

[0045] In a fifth aspect, the present invention provides a computer program product comprising computer instructions for causing a computer to execute the intelligent imaging scanning method of the first aspect or any corresponding embodiment thereof.

[0046] It should be noted that the intelligent imaging scanning system, computer device, computer-readable storage medium, and computer program product provided herein correspond to the intelligent imaging scanning method described above. Therefore, for the beneficial effects of the intelligent imaging scanning system, computer device, computer-readable storage medium, and computer program product, please refer to the description of the corresponding beneficial effects of the intelligent imaging scanning method above and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 is a flow chart of an intelligent imaging scanning method according to an embodiment of the present invention;

[0049] Figure 2 is a structural block diagram of an intelligent imaging scanning system according to an embodiment of the present invention;

[0050] Figure 3 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0052] Take CT scans as an example. Currently, the requirements for scanning technicians are becoming increasingly stringent. Because CT scans involve ionizing radiation, the selection of scanning sequences is particularly crucial. Scanning sequences must not only be accurate but also kept to a minimum, meeting diagnostic requirements. Scanning sequence selection is typically adjusted in real time by the scanning technician based on the examination images. An incorrectly selected scanning sequence can not only render the scanned images inconclusive but also increase the patient's ionizing radiation exposure. Inappropriately selected scanning sequences can lead to unclear diagnoses or even missed diagnoses. Therefore, the expertise and experience of the scanning technician are crucial. However, even experienced scanning technicians can make mistakes in operation, select the wrong scanning sequence, or simply lack the knowledge to select the appropriate scanning sequence for unusual imaging types.

[0053] In view of this, according to an embodiment of the present invention, an embodiment of an intelligent imaging scanning method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0054] In this embodiment, an intelligent imaging scanning method is provided, which can be executed by an intelligent imaging scanning system, a server, a terminal, a mobile terminal, and other devices. Figure 1 FIG. 1 is a flow chart of an intelligent imaging scanning method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0055] Step S101 obtains the scan results of the individual to be scanned under the current scan protocol. The current scan protocol is determined based on the information in the patient's checklist. In this embodiment, a CT scan is used as an example. For example, if the checklist indicates that a "chest CT plain scan" is required, the scan protocol corresponding to this "chest CT plain scan" is directly retrieved and used as the current scan protocol. Scans are then performed sequentially according to the scan sequence in the scan protocol, and the results of each scan are sent to a pre-established imaging disease recognition model or server device.

[0056] In step S102, the disease recognition result is determined based on the scan results and a pre-established imaging disease recognition model. The imaging disease recognition model can identify and diagnose the scan results in real time. The imaging disease recognition model used in this embodiment is trained with a large amount of valid data, enabling it to accurately and quickly identify various anatomical organs, diagnose them, and generate a text report.

[0057] Step S103 , when the disease identification result indicates that the disease exists, a target scanning protocol corresponding to the disease type in the disease identification result is retrieved from a pre-established scanning protocol database; the scanning protocol database includes scanning protocols corresponding to different disease types.

[0058] Specifically, the disease type is extracted from the disease recognition results predicted by the imaging disease recognition model, and then the scan protocol database is searched for the corresponding scan protocol based on the disease type. If the disease recognition result indicates that the disease is not present, the scan result is directly output, and the scan is determined to be complete.

[0059] The scanning protocol database in this embodiment can be established using the following structure:

[0060] The first-level directory of the scanning protocol database is classified by equipment brand name and equipment model, such as GERevolution APEX CT 256-row 512-layer flagship model and Siemens Naeotom Alpha.Peak flagship model.

[0061] The secondary directory of the scanning protocol database is classified by human tissue and / or structure, such as chest, abdomen, spine, blood vessels, various organs, etc.

[0062] The three-level directory of the scan protocol database is categorized by CT imaging examination, such as plain chest CT, enhanced abdominal CT, and aortic CT angiography. Each CT imaging examination corresponds to a scan sequence. For example, to perform an enhanced abdominal CT scan, the scan protocol specifies the order and timing of multiple scan sequences. These sequences perform different types of image acquisition based on different requirements. These scan sequences include: scout scan sequence, plain scan phase scan sequence, monitoring scan sequence, detection ROI curve trigger sequence, enhanced arterial phase scan sequence, enhanced portal venous phase scan sequence, and so on.

[0063] In addition, the scanning protocol database is also linked to the disease-specific database. Each disease type leads to the corresponding CT imaging examination items, such as chest CT plain scan → pulmonary bullae, chest CT plain scan → routine chest examination (used to examine diseases including pneumonia, lung cancer, etc.), abdominal CT enhancement → hepatic hemangioma, etc.

[0064] The scan protocol database also includes parameter information and guidance for different CT imaging procedures and disease types, including different equipment parameters. This information includes original image reconstruction information (conventional slice thickness, slice interval, thin slice thickness, slice interval), reconstruction algorithm, tube voltage selection, tube current selection, other parameters (collimation width), window technology, enhanced imaging, startup method, and contrast agent regimen. Guidance information also includes examination position and scanning respiratory coordination (inspiratory breath-hold, inspiratory-expiratory breath-hold).

[0065] For example, when the type of device used is determined and it is determined that the patient may have a hepatic hemangioma, the scan protocol database will generate a corresponding target scan protocol from the scan protocol database based on the disease type, namely "hepatic hemangioma". For example, according to "hepatic hemangioma", it is directed to "abdominal CT enhancement", and according to "abdominal CT enhancement", the scan sequence of "abdominal CT enhancement → hepatic hemangioma" is directly retrieved. This scan sequence includes a positioning image scan sequence, a plain scan phase scan sequence, a monitoring layer scan sequence, a detection ROI curve trigger sequence, an enhanced arterial phase scan sequence, an enhanced portal venous phase scan sequence, and the scan parameters corresponding to each scan sequence. At this time, the device type, CT imaging examination items, scan sequence, scan parameters, etc. are encapsulated as the target scan protocol.

[0066] Step S104: Based on the current scanning protocol and the target scanning protocol, it is determined whether a supplementary scan is required.

[0067] Compare the scan information in the current scan protocol with the scan information in the target scan protocol, such as comparing the CT imaging examination items, scan sequence, scan parameters, etc., to determine whether a supplementary scan is needed.

[0068] Step S105: If it is determined that a supplementary scan is required, a supplementary scan protocol is generated.

[0069] Step S106: Continue scanning based on the supplementary scanning protocol until the scan is completed.

[0070] The supplementary scanning protocol may include CT imaging examination items that require supplementary scanning, or one or more scanning sequences, etc.

[0071] For example, consider extracting the lesion keyword "hemangioma liver cancer" from the disease recognition results of the imaging disease recognition model. Based on the keyword "hemangioma liver cancer," the scan protocol database is searched for the corresponding scan protocol for "abdominal CT enhancement → hemangioma liver cancer." The retrieved scan protocol is compared with the scan protocols already examined. After comparison, any duplicate scan protocols are deleted, generating the remaining unscanned protocols (the supplemental scan protocols). Finally, the examination is completed according to the scan protocols.

[0072] In typical scanning procedures, the scanning technician typically adjusts the scanning protocol in real time based on the imaging findings during the scan to optimize the best protocol for lesion diagnosis. For example, when a patient undergoes an enhanced abdominal scan, after completing the plain scan, enhanced arterial phase, and portal venous phase scans, the scanning technician will make a preliminary diagnosis based on the image content. If the patient's liver imaging reveals imaging findings resembling a hemangioma or liver cancer, the scanning technician will add a liver delay scan protocol and calculate the delay time to perform the scan, thereby distinguishing between hemangiomas and liver cancer. Studies have found that scanning technicians frequently adjust the scanning protocol based on real-time images, which can easily lead to inaccurate adjustments. For example, during a plain chest CT scan, the scanning technician needs to observe the scan image in real time to identify imaging findings characteristic of a disease. For example, if a scan shows patchy ground-glass opacities or slightly hyperdensitivity with blurred edges on the dorsal aspect of the lower lung, the technician will manually add a prone chest CT plain scan protocol to differentiate between chest effusion, interstitial lung disease, or pneumonia. When the imaging features of the lesion disappear on the image after the scan is completed, the lesion is confirmed to be a chest accumulation effect. Definitive imaging diagnosis can not only avoid the patient having to undergo other examinations for differential diagnosis, but also allow the clinician to directly provide an accurate treatment plan based on the confirmed results. However, during the scanning process, it is extremely difficult to complete the above operations. The scanning process leaves little time for the scanning technician to make judgments. In addition, the scanning technician is required to have a lot of imaging diagnosis knowledge, understand the lesions corresponding to the imaging features, and the scanning methods and scanning protocol contents for distinguishing lesions with the same type of imaging features.

[0073] For example, a space-occupying lesion on the kidney is detected during an enhanced abdominal CT scan. During the arterial phase, this lesion often exhibits uneven and pronounced enhancement, with some vascular structures visible within the lesion. During the portal venous phase, the lesion's enhancement rapidly decreases, exhibiting a "quick in, quick out" pattern. To clarify the nature of the lesion, a delayed scan protocol of 3-5 minutes, or even longer, should be performed. During the delayed phase, the lesion's density decreases further, contrasting more clearly with the surrounding renal parenchyma. Because the imaging features are generally consistent with the imaging manifestations of this lesion and renal cell carcinoma, and because enhanced abdominal CT scans involve numerous dissected organs, relying on the scanning technician to visually diagnose all lesions without missing any, and then editing the corresponding lesion protocol for scanning, this process is difficult to complete. This process also requires the contrast agent to maintain its bolus form within the patient's body, which is a long-term requirement.

[0074] The intelligent imaging scanning method proposed in this embodiment combines the scanning protocol database with the imaging disease recognition model. It can predict the type of disease that the patient may have based on the scanning results, and generate a target scanning protocol that needs to be scanned based on the disease type. According to the current scanning protocol and the target scanning protocol, it is determined whether a supplementary scan is needed to obtain more comprehensive imaging information. According to the intelligent imaging scanning method provided in this embodiment, when performing CT scanning or MRI scanning, there is no need to over-rely on the professional knowledge and experience of the scanning technician. That is, the scanning protocol can be intelligently and adaptively adjusted for different patients, different disease states, and different scanning sites to ensure that each patient can obtain the most comprehensive and accurate scanning results, and there will be no missed scans and the need for supplementary scans. In addition, the disease recognition results determined according to the imaging disease recognition model can also assist medical staff in diagnosis to prevent missed judgments and misjudgments.

[0075] In addition, the system only generates a supplemental scan protocol when it determines that a supplemental scan is needed. Supplemental scan protocols are generated based on the patient's specific real-time situation, with personalized scan protocols generated based on the patient's physical condition, disease type, and scanning needs. When this imaging scanning method is applied to CT scans, it can effectively avoid using a uniform high-dose scanning protocol for all patients. In other words, the intelligent imaging scanning method provided by the present invention can also avoid overscanning, reduce radiation dose usage, and thus reduce ionizing radiation exposure to patients, thereby increasing patient safety.

[0076] In some optional embodiments, the current scanning protocol includes multiple scanning sequences, and the disease identification result is determined based on the scanning results and a pre-established imaging disease identification model, including:

[0077] The scanning results corresponding to each scanning sequence are sequentially input into the imaging disease recognition model to obtain a one-to-one corresponding prediction result for each scanning sequence;

[0078] Based on all the prediction results, the disease identification result is determined.

[0079] For example, it is necessary to sequentially execute the plain phase scanning sequence, the monitoring layer scanning sequence, the detection ROI curve trigger sequence, the enhanced arterial phase scanning sequence, and the enhanced portal venous phase scanning sequence. The plain phase scanning sequence begins, and during the scanning process, the imaging disease recognition model initiates real-time diagnosis of the scanning results. When the abdominal aorta is located, the monitoring layer scanning sequence begins. On the monitoring layer image, the ROI monitoring curve monitoring point is placed within the abdominal aorta, and the ROI curve trigger sequence scan begins. When the threshold within the monitoring area reaches the trigger condition, the enhanced arterial phase scan begins. During the sequence scanning process, the imaging disease recognition model performs real-time predictions based on the real-time scanned images. After the enhanced arterial phase scan is completed, the enhanced portal venous phase sequence begins scanning after a set delay time. During the scanning process, the imaging disease recognition model again initiates real-time predictions of the scanning results, and finally combines all the prediction results to output the disease recognition results.

[0080] In this embodiment, the final disease identification result is comprehensively determined based on the prediction results of all scanning sequences, which can effectively increase the prediction accuracy, thereby improving the accuracy and comprehensiveness of the scanning results.

[0081] Furthermore, after the supplemental scan, the imaging disease recognition model can be used for further predictions. For example, based on the imaging features of the scan, the disease identification result may indicate suspected hepatic hemangioma or liver cancer, requiring a delayed scan for further identification. Following the generated supplemental scan protocol, another scan is performed to obtain the final scan results. At this point, the imaging disease recognition model re-predicts the structures in the real-time scan, resulting in a final diagnosis of a high probability of liver cancer, further increasing the accuracy of the prediction results.

[0082] In some optional implementations, the scanning protocol database further includes scanning protocols corresponding to different scanning information. The current scanning protocol is determined by the following steps:

[0083] Obtain a checklist of individuals to be scanned;

[0084] Identify scanned information in the checklist;

[0085] A current scan protocol for the individual to be scanned is determined based on the scan information and a scan protocol database.

[0086] For example, if the checklist requires a "chest CT plain scan," the scan information, i.e., "chest CT plain scan," is extracted through text recognition. Based on the scan information, the scan protocol corresponding to "chest CT plain scan" is directly retrieved and used as the current scan protocol.

[0087] In this embodiment, the corresponding scanning protocol can be automatically determined based on the patient's examination list, and the scan can be automatically completed, which reduces the dependence on the professional knowledge and experience of the scanning technician and improves the efficiency and accuracy of the scan.

[0088] In some optional embodiments, the current scanning protocol includes a scout image scanning sequence, and the method further includes:

[0089] The scout image scan is performed based on the scout image scan sequence to obtain a scout image scan result. The scout image scan in this embodiment is often to take a low-dose whole-body scan image to help technicians accurately locate the scan area.

[0090] The positioning image scan results are input into the imaging disease recognition model to determine whether there are possible abnormal areas.

[0091] When it is determined that there is a possible abnormal area, the scanning starting point and scanning range of other scanning sequences in the current scanning protocol are updated;

[0092] Other scanning sequences are scanned based on the updated scanning starting point and scanning range to obtain corresponding scanning results.

[0093] That is, before officially starting the scan, you can first perform a positioning image scan based on the positioning image scan sequence. At this time, the imaging disease recognition model is started for the first time, and a preliminary diagnosis is made based on the two-dimensional data of the positioning image to identify and confirm whether there are obvious lesions and identify tissues with obvious differences in CT values such as bones and lungs to automatically identify the scan boundary according to the scanning protocol range. That is, the position, morphology and spatial distribution of the lesion are identified and located, and then the scanning starting point and scanning range of other scanning sequences that need to be scanned are re-determined.

[0094] In this embodiment, a locator scan is performed before the actual scan to determine the accurate scan starting point and scanning range. This can initially ensure the comprehensiveness of the scanning range. On this basis, a supplementary scan can further improve the completeness of the scan, achieving a double guarantee of the comprehensiveness of the scanning range and preventing missed scans and the need for supplementary scans.

[0095] In some optional embodiments, the imaging disease recognition model is established by the following steps:

[0096] Step a1, obtain a data set, which includes standard images and lesion images. Taking CT scans as an example, the data source can include hospital CT image databases or public data sets (such as LIDC-IDRI) that require ethical review and patient authorization. The data content includes: standard DICOM format CT images of each examination site and DICOM format CT images of clearly confirmed lesions in the corresponding site (such as emphysema, pneumonia, pulmonary accumulation effect, etc. on chest CT). DICOM format image data includes multi-morphological data such as transverse, coronal, parasitic, and three-dimensional imaging.

[0097] Step a2: Construct an initial network structure. This initial network structure is a multimodal model that combines a convolutional neural network, a graph convolutional network, and a multi-task learning model. This initial network structure can adopt a multimodal approach that combines a convolutional neural network (CNN) with a graph convolutional network (GCN) and a multi-task learning model (MTL), such as U-Net for image segmentation and ResNet for classification.

[0098] Specifically, the convolutional neural network (CNN) includes: a convolutional layer for extracting local features (such as the edge of lung nodules and vascular textures); a pooling layer for compressing feature dimensions and retaining key information (such as significant tumor areas); a fully connected layer for integrating global features and outputting classification probabilities (such as benign and malignant discrimination); a feature extraction module for automatically identifying the morphological features of lesions (such as size, density, and edge burrs) and functional parameters (such as dynamic enhancement curves). The graph convolutional network (GCN) includes: a lesion detection module for locating abnormal areas (such as lung nodules and coronary artery plaques). Through the graph structure characteristics of GCN, lesions can be effectively and accurately located, with a detection rate of up to 99.2% (nodules below 5mm). The multi-task learning model includes: a quantitative analysis module for measuring lesion volume, CT value, and hemodynamic parameters (such as FFRct), and the error rate can be kept below 5%.

[0099] Step a3: Train the initial network structure based on the data set to obtain an image disease recognition model.

[0100] In this embodiment, the convolutional neural network can effectively process CT data (such as thin-layer scans of the lungs) and capture the spatial distribution characteristics of lesions (such as the spicule sign of lung cancer). The graph convolutional network (GCN) can effectively construct graph structures (such as vascular networks, spatial correlations between lesions and surrounding tissues), establish spatial topological relationships, and enhance the ability to capture global features. Coupled with the multi-task learning model (MTL), it can handle multiple related tasks at the same time, helping the model to perform quantitative analysis while classifying diseases, thereby improving the overall model performance. Multimodal learning can further enhance the generalization ability of the model by integrating information from different sources. Comparing and aligning the processed images with the database can help improve the accuracy of diagnosis. The use of a multimodal model can effectively improve the recognition accuracy of the final imaging disease recognition model, thereby effectively improving the accuracy and comprehensiveness of the scanning range.

[0101] In some optional implementations, after completing the scan, the method further includes:

[0102] The disease identification results corresponding to the individual to be scanned and the relevant scanning protocols are associated with the scanning protocol database.

[0103] After the scan is completed, the overall scan protocol and diagnostic results can be linked to the scan protocol database, and connected through the RIS (a radiology information system) and HIS (a hospital information system) databases to generate a customized scan sequence specifically for this patient. The next time this patient comes for a follow-up examination, the appropriate scan protocol for this patient's lesion can be directly called up. Moreover, a better scan protocol can be matched according to the equipment used by the patient. For example, high-end CT equipment can perform energy spectrum scanning for this patient. Energy spectrum scanning can distinguish between hepatoblastoma, hemangioma and metastatic tumors through iodine-based image and water-based image separation technology. For example, rich blood supply tumors (such as hemangiomas) are more significantly enhanced in low-keV single-energy images, which is more effective in diagnosing the patient's lesions.

[0104] All this data can be stored in a local database or uploaded to a cloud storage, and the scanning protocol database can be improved through more and more classified scanning data.

[0105] The database management system also enables storage, retrieval, backup, and sharing, while supporting remote data addition and access. Furthermore, database permission management can be added to prevent arbitrary parameter modification and misoperation. Scan protocol databases can also be integrated with multimodal data to build a cross-modality database combining PET-CT, MRI, or ultrasound data.

[0106] In addition to helping scanning technicians accurately adjust the scanning protocol according to different patient conditions to accurately complete the examination process, the present invention can also store the entire process in a database and connect it with RIS and HIS. In this way, both the reporting doctor and the clinician can know the adjusted scanning protocol and the target lesion, and can observe it accurately and quickly.

[0107] The specific scanning process can be as follows: First, the patient lies on the examination bed. Based on the scan area, the scanning technician manually confirms the scan baseline. The system then automatically generates the current scan protocol based on the examination information on the checklist and scans the positioning image according to the positioning baseline. Based on the scan positioning image, the selected scan protocol automatically selects the scan starting point for each scan sequence and begins scanning. During the scanning process of each scan sequence, the imaging disease recognition model performs real-time disease prediction and ultimately determines whether to trigger scan protocol adjustment. If adjustment is required, the scan continues using the adjusted supplementary scan protocol until the scan is complete. Finally, all data is recorded and stored.

[0108] This embodiment also provides an intelligent imaging scanning system for implementing the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the systems described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0109] This embodiment provides an intelligent imaging scanning system. Figure 2 As shown, the system includes:

[0110] An acquisition module 201 is used to acquire a scan result of the subject to be scanned under the current scan protocol;

[0111] The imaging diagnosis auxiliary module 202 is used to determine the disease identification result based on the scanning results and a pre-established imaging disease identification model; it is also used to input the scanning results corresponding to each scanning sequence into the imaging disease identification model in sequence to obtain a one-to-one corresponding prediction result for each scanning sequence; and determine the disease identification result based on all the prediction results.

[0112] The retrieval module 203 is configured to retrieve a target scanning protocol corresponding to the disease type in the disease identification result from a pre-established scanning protocol database when the disease identification result indicates the presence of the disease; the scanning protocol database includes scanning protocols corresponding to different disease types;

[0113] A determination module 204 is configured to determine whether a supplementary scan is required based on the current scan protocol and the target scan protocol;

[0114] The generation module 205 is used to generate a supplementary scan protocol when it is determined that a supplementary scan is required; it is also used to obtain a checklist of the individual to be scanned; identify the scan information in the checklist; and determine the current scan protocol of the individual to be scanned based on the scan information and the scan protocol database.

[0115] The scanning module 206 is configured to continue scanning based on the supplementary scanning protocol until the scanning is completed.

[0116] In some optional embodiments, the system further includes:

[0117] The scanning module is used to perform positioning image scanning based on the positioning image scanning sequence to obtain positioning image scanning results; input the positioning image scanning results into the image disease recognition model to determine whether there is a possible abnormal area; when it is determined that there is a possible abnormal area, update the scanning starting point and scanning range of other scanning sequences in the current scanning protocol; other scanning sequences are scanned based on the updated scanning starting point and scanning range to obtain corresponding scanning results.

[0118] The model building module is used to obtain a data set, which includes standard images and pathological images; construct an initial network structure, where the initial network structure is a multimodal model that combines a convolutional neural network, a graph convolutional network, and a multi-task learning model; and train the initial network structure based on the data set to obtain an image disease recognition model.

[0119] The association module is used to associate the disease identification results corresponding to the individual to be scanned and the relevant scanning protocols with the scanning protocol database.

[0120] The intelligent imaging scanning system in this embodiment is presented in the form of functional units, where the units refer to ASIC circuits, processors and memories that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0121] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0122] The embodiment of the present invention also provides a computer device having the above Figure 2 The intelligent imaging scanning system shown.

[0123] See also Figure 3 , Figure 3 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 3As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of a GUI on an external input / output device (such as, a display device coupled to an interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 3 A processor 10 is taken as an example.

[0124] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0125] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0126] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0127] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0128] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0129] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0130] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0131] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. An intelligent imaging scanning method, characterized in that: The method comprises: Obtain the scan results of the individual to be scanned under the current scanning protocol; Determining a disease recognition result based on the scan result and a pre-established image disease recognition model; If the disease identification result indicates that the disease exists, a target scanning protocol corresponding to the disease type in the disease identification result is retrieved from a pre-established scanning protocol database; the scanning protocol database includes scanning protocols corresponding to different disease types; Determining whether a supplementary scan is required based on the current scan protocol and the target scan protocol; If it is determined that a supplementary scan is required, generating a supplementary scan protocol; Scanning continues based on the supplemental scanning protocol until the scan is complete.

2. The method according to claim 1, characterized in that The current scanning protocol includes a plurality of scanning sequences, and determining a disease identification result based on the scanning result and a pre-established imaging disease identification model includes: The scanning results corresponding to each scanning sequence are sequentially input into the imaging disease recognition model to obtain a one-to-one corresponding prediction result for each scanning sequence; Based on all the prediction results, the disease identification result is determined.

3. The method according to claim 1, characterized in that The scanning protocol database also includes scanning protocols corresponding to different scanning information. The current scanning protocol is determined by the following steps: Obtaining a checklist of the individual to be scanned; identifying scan information in the checklist; The current scanning protocol of the individual to be scanned is determined based on the scanning information and the scanning protocol database.

4. The method according to claim 1, wherein The current scanning protocol includes a scout image scanning sequence, and the method further includes: Performing a scout image scan based on the scout image scan sequence to obtain a scout image scan result; Inputting the scout image scan result into the image disease recognition model to determine whether there is a possible abnormal area; When it is determined that the possible abnormal area exists, updating the scanning starting point and scanning range of other scanning sequences in the current scanning protocol; The other scanning sequences perform scanning based on the updated scanning starting point and the scanning range to obtain the corresponding scanning results.

5. The method according to claim 1, wherein The image disease recognition model is established by the following steps: Acquiring a data set, wherein the data set includes a standard image and a lesion image; Constructing an initial network structure, wherein the initial network structure is a multimodal model that combines a convolutional neural network, a graph convolutional network, and a multi-task learning model; The initial network structure is trained based on the data set to obtain the image disease recognition model.

6. The method according to claim 1, characterized in that After the scanning is completed, the method further includes: The disease identification result and related scanning protocol corresponding to the individual to be scanned are associated with the scanning protocol database.

7. An intelligent imaging scanning system, characterized in that: The system comprises: An acquisition module is used to obtain the scan result of the individual to be scanned under the current scanning protocol; An image diagnosis auxiliary module, configured to determine a disease identification result based on the scan result and a pre-established image disease identification model; a retrieval module, configured to retrieve, when the disease identification result indicates the presence of a disease, a target scanning protocol corresponding to the disease type in the disease identification result from a pre-established scanning protocol database; the scanning protocol database includes scanning protocols corresponding to different disease types; A judgment module, configured to judge whether a supplementary scan is required based on the current scan protocol and the target scan protocol; A generating module, configured to generate a supplementary scanning protocol when it is determined that a supplementary scanning is required; The scanning module is configured to continue scanning based on the supplementary scanning protocol until the scanning is completed.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the intelligent imaging scanning method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the intelligent imaging scanning method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to enable a computer to execute the intelligent imaging scanning method according to any one of claims 1 to 6.

Images