Medical image scheduling method, scheduling system and storage medium implemented by computer

The medical image scheduling method implemented by computer uses artificial intelligence to make preliminary judgments and sort medical images, and combines clinician diagnosis results and data acquisition time to solve the problem of inaccurate case priorities in the existing system, ensuring the case Timely processing and efficient sorting.

CN112466439BActive Publication Date: 2024-12-03SHENZHEN KEYA MEDICAL TECH CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011412540.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-01-23
Filing Date
2018-03-09
Publication Date
2024-12-03
Estimated Expiration
2038-03-09

AI Technical Summary

Technical Problem

The existing medical image scheduling system has inaccurate and unreliable problems in case prioritization, which leads to slowing down case processing and even delaying the optimal treatment time window, especially when the patient characterization is not obvious. Making a misjudgment.

Method used

Using the medical image scheduling method implemented by computer, the medical image is initially judged through artificial intelligence, image-related data is obtained, and priority parameters are obtained. The clinician's diagnosis results and data acquisition time are accurately sorted, and the queue is dynamically adjusted. Sequence to meet the real processing needs.

Benefits of technology

The timely handling of important cases of emergency and important cases has been achieved, the accuracy and efficiency of case handling have been improved, the workload allocation of imaging doctors has been optimized, and processing delays have been avoided due to improper allocation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112466439B_ABST
    Figure CN112466439B_ABST
Patent Text Reader

Abstract

The present disclosure relates to a medical image scheduling method, scheduling system and storage medium implemented by a computer. The method comprises the following steps: obtaining medical image related data; using a processor to process the medical image related data to at least obtain priority parameters of the medical image related data; using a processor to sort the medical image related data at least according to the priority parameters of the medical image related data; and using a processor to present a queue of medical image related data and its processing results according to the sorting. The method makes a preliminary judgment on the medical image based on artificial intelligence and optimizes the scheduling accordingly. The medical image scheduling method is accurate, reliable and rapid. When necessary, it can also combine the judgment results of artificial intelligence, the attribute information of the medical image (such as data acquisition time) and the diagnosis results of the clinician to accurately sort the cases.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application is a divisional application of the Chinese invention patent application with application number 201810194503.5, application date March 9, 2018, and invention name “Medical image scheduling method, scheduling system and storage medium implemented by computer”. Technical Field

[0002] The present invention relates to an artificial intelligence-based workflow optimization method, and more particularly to a medical image scheduling method implemented by a computer. The present invention also relates to a medical image scheduling system and a computer-readable storage medium storing information capable of executing the method. Background Art

[0003] With the continuous development of technology, medical images are playing an indispensable role in the process of clinical diagnosis. After being collected, most of these images need to be reviewed and annotated by imaging doctors (such as radiologists) and complete pathological reports for reference by clinicians. Since each imaging doctor needs to process a large number of images a day, most hospitals or Internet-based remote diagnosis platforms will use a queuing system to manage these images, and then the imaging doctor will select the medical images of the case for processing according to the priority in the queue.

[0004] In general, traditional systems will sort images according to the time of image acquisition, so that each patient's image will not stay in the queue for too long, that is, the ranking of the first patient is higher than that of the later patient. More advanced systems should comprehensively consider the data acquisition time, the time urgency of the disease, and the workload of doctors to reasonably schedule and allocate medical-related data. For example, cerebral hemorrhage and cerebral thrombosis are two common acute diseases. Although both are brain blood diseases, the treatment methods of these two diseases are completely different. If the doctor can make the right judgment within a few hours of the onset of the disease and provide immediate assistance, the danger of the disease will be greatly reduced and the patient's recovery will be guaranteed. For such images, the queue management system needs to queue up the images of these cases and put them at the top of the queue, marking them as emergency cases to remind the radiologists. If an emergency case that needs to be handled immediately is found in the queue, the radiologist needs to put down the non-emergency work in his hands and deal with it immediately.

[0005] It should be pointed out that in the current system that allows cases to be sorted by urgency, the priority of the medical images of the case is derived from the priority of its condition, and the priority of the condition is determined only by the clinician of the case based on the patient's manifestation. For this system, it cannot be ruled out that the clinician may make a misjudgment when the patient's manifestation is not obvious, resulting in underestimation of the priority of the case, thereby slowing down the processing speed of the case and delaying the optimal treatment time window. Furthermore, clinicians often need to refer to the processing results of the medical images of the case by radiologists to make accurate judgments. Sometimes when sorting the medical images of the case, an accurate medical diagnosis has not been made for the patient of the case, resulting in a lack of accurate priority basis for sorting. In addition, clinicians face heavy diagnostic and treatment work. It is risky, inefficient, and not intelligent enough to rely too much on their artificial diagnosis results of the case to sort the medical images of the case. The above is just an example. There are many factors in the current system that allows cases to be sorted by urgency that will lead to inaccurate or unreliable case priorities, slow down the processing speed of cases (including their medical images), and even delay the optimal treatment time window. Summary of the invention

[0006] In view of the above problems existing in the prior art, the purpose of the present invention is to provide a medical image scheduling method implemented by a computer that can ensure that urgent and important cases can be processed early and in a timely manner. The method makes a preliminary judgment on medical images based on artificial intelligence and optimizes the scheduling accordingly. This medical image scheduling method is accurate, reliable and rapid. When necessary, it can also combine the judgment results of artificial intelligence, the attribute information of medical images (such as data acquisition time) and the diagnosis results of clinicians to accurately sort the cases. Furthermore, the method can also provide dynamic sorting in near real time with the introduction of new medical images, so that the queue order of the presented medical images conforms to and meets the actual processing needs of all cases at that time.

[0007] In order to achieve the above object, an embodiment of the present invention provides a medical image scheduling method in one aspect, which is implemented by a computer and is characterized in that the medical image scheduling method comprises the following steps:

[0008] Obtain medical image-related data;

[0009] Using a processor, processing the medical image related data to obtain at least a priority parameter of the medical image related data;

[0010] sorting, using a processor, the medical image related data according to at least a priority parameter of the medical image related data; and

[0011] A processor is used to present a queue of medical image-related data and processing results thereof according to the sorting.

[0012] Preferably, the step of processing the medical image related data to obtain the priority parameter of the medical image related data comprises:

[0013] Classifying the medical image-related data according to modality and imaging site;

[0014] According to the classification result, a corresponding image analyzer is used to analyze the medical image to at least obtain a priority parameter of medical image related data.

[0015] Preferably, the step of classifying the medical image-related data according to modality and imaging site is performed at least based on image features of the medical image.

[0016] Preferably, the classification of the medical image-related data according to modality and imaging site is achieved by an image retrieval algorithm and / or an image classification algorithm.

[0017] Preferably, the image analyzer comprises a plurality of image analyzers, each image analyzer being designed to perform calculations for a specific modality and imaging site.

[0018] Preferably, the image analyzer is implemented using one or more algorithms of deep neural network, random forest, gradient boosting decision tree.

[0019] Preferably, the result of processing the medical image related data includes the diagnosis result, related area and priority parameter of the medical image.

[0020] Preferably, the priority parameter of the medical image related data is obtained according to a predetermined rule based on the diagnostic condition of the medical image.

[0021] Preferably, the predetermined rule includes: after the medical image related data enters the queue, its priority parameter increases with time at a certain rate, and the certain rate is based on the initial priority parameter of the medical image related data and its specified processing time.

[0022] Preferably, the priority parameter of the medical image related data is obtained by using a regression algorithm based on image features of the medical image.

[0023] Preferably, the step of processing the medical image related data to obtain at least a priority parameter of the medical image related data comprises:

[0024] When there is a problem with the diagnosis result obtained by the processor, the diagnosis result is corrected.

[0025] Preferably, the correction comprises:

[0026] sending the medical image-related data to a first doctor and obtaining the first doctor's diagnosis result; comparing the diagnosis result obtained by the processor with the diagnosis result of the first doctor; and

[0027] Based on the comparison results, the diagnosis results, relevant areas and priority parameters of the medical image related data are obtained.

[0028] Preferably, the step of obtaining the diagnosis result, the relevant area and the priority parameter of the medical image related data based on the comparison result comprises:

[0029] When the comparison results are inconsistent, send the two diagnosis results to other doctors; and

[0030] Based on the doctor's comprehensive diagnosis results, the diagnosis results, relevant areas and priority parameters of medical image-related data are obtained.

[0031] Preferably, the medical image scheduling method comprises:

[0032] The processor is used to filter the acquired medical image-related data in response to acquiring the medical image-related data, and the filtered processable data is provided to the processor for subsequent processing.

[0033] Preferably, the step of sorting the medical image related data at least according to the priority parameters of the medical image related data comprises:

[0034] The medical image related data are sorted according to priority parameters of the medical image related data and also according to at least one of a doctor's diagnosis result and a data acquisition time.

[0035] Preferably, the medical image scheduling method further includes:

[0036] In response to applying a corresponding image analyzer to analyze the medical image according to the classification result, the analysis result of the image analyzer is comprehensively analyzed to obtain priority parameters and diagnosis results of medical image related data.

[0037] Preferably, the comprehensive analysis step includes:

[0038] When the user has different medical image-related data, the analysis results of the different data by the image analyzer are comprehensively analyzed.

[0039] Preferably, the comprehensive analysis step includes:

[0040] A comprehensive analysis is performed on the diagnosis results of the medical image related data, the image features of the medical image, and the attribute information of the patient of the medical image.

[0041] According to a second aspect of the present invention, a computer-readable storage medium is provided, which stores computer-executable instructions. When the computer-executable instructions are executed by a processor, any one of the above-mentioned medical image scheduling methods implemented by a computer is implemented.

[0042] According to a third aspect of the present invention, there is provided a medical image scheduling system, which is communicatively connected to an image acquisition unit to obtain medical image related data therefrom, characterized in that the system comprises:

[0043] a data processing unit configured to process the acquired medical image-related data to obtain at least a priority parameter of the medical image-related data;

[0044] A data storage scheduling unit configured to sort the medical image related data at least according to a priority parameter of the medical image related data; and

[0045] The data browsing section is configured to present a queue of medical image-related data and processing results thereof according to sorting.

[0046] Preferably, the data processing unit includes:

[0047] an image sorter configured to classify the medical image-related data with respect to modality and imaging site;

[0048] An image analysis module, including a plurality of image analyzers, each image analyzer being designed to perform calculations for a specific modality and imaging site;

[0049] Wherein, the data processing unit uses a corresponding image analyzer to analyze the medical image according to the classification result of the image sorter, so as to at least obtain the priority parameters of the medical image related data.

[0050] Preferably, the classification of the medical image-related data with respect to modality and imaging region is performed at least based on image features of the medical image.

[0051] Preferably, the classification of the medical image-related data according to modality and imaging site is achieved by an image retrieval algorithm and / or an image classification algorithm.

[0052] Preferably, the image analyzer is implemented using one or more algorithms of deep neural network, random forest, gradient boosting decision tree.

[0053] Preferably, the result of processing the medical image related data includes the diagnosis result, related area and priority parameter of the medical image.

[0054] Preferably, the data processing unit is further configured to obtain a priority parameter of the medical image related data based on the diagnosis result of the medical image according to a predetermined rule.

[0055] Preferably, the predetermined rule includes: after the medical image related data enters the queue, its priority parameter increases with time at a certain rate, and the certain rate is based on the initial priority parameter of the medical image related data and its specified processing time.

[0056] Preferably, the image analyzer is configured to obtain priority parameters of medical image related data using a regression algorithm based on image features of the medical image.

[0057] Preferably, the data processing unit further comprises a comprehensive analysis module, which is configured to correct the obtained diagnostic result when there is a problem with the diagnostic result obtained by the image analysis module.

[0058] Preferably, the comprehensive analysis module is further configured as follows:

[0059] Sending the medical image-related data to the first doctor and obtaining the first doctor's diagnosis result;

[0060] comparing the diagnosis result obtained by the image analysis module with the diagnosis result of the first doctor; and

[0061] Based on the comparison results, the diagnosis results, relevant areas and priority parameters of the medical image related data are obtained.

[0062] Preferably, the comprehensive analysis module is further configured as follows:

[0063] When the comparison results are inconsistent, send the two diagnosis results to other doctors; and

[0064] Based on the doctor's comprehensive diagnosis results, the diagnosis results, relevant areas and priority parameters of medical image-related data are obtained.

[0065] Preferably, the medical image scheduling system comprises an image filter configured to:

[0066] In response to acquiring the medical image related data, the acquired medical image related data is screened, and the screened processable data is provided to the image sorter for subsequent processing.

[0067] Preferably, the data storage scheduling unit is further configured as follows:

[0068] The medical image related data are sorted according to priority parameters of the medical image related data and also according to at least one of a doctor's diagnosis result and a data acquisition time.

[0069] Preferably, the comprehensive analysis module is further configured as:

[0070] In response to the data processing unit applying a corresponding image analyzer to analyze the medical image according to the classification result, the analysis result of the image analyzer is comprehensively analyzed to obtain at least a priority parameter of the medical image related data.

[0071] Preferably, the comprehensive analysis module is further configured as:

[0072] When the user has different medical image-related data, the analysis results of the different data by the image analyzer are comprehensively analyzed.

[0073] Preferably, the comprehensive analysis module is further configured as:

[0074] A comprehensive analysis is performed on the diagnosis results of the medical image related data, the image features of the medical image, and the attribute information of the patient of the medical image.

[0075] The computer-implemented medical image scheduling method and system provided by the present invention can be used by a unified medical image platform system to transmit, store, pre-diagnose, sort, schedule, and record the results of medical images waiting for radiologists to browse. For non-urgent cases, the system will also comprehensively consider the workload of radiologists in the process of scheduling and allocating cases. In this way, on the one hand, by using artificial intelligence to recommend the order of processing medical images to radiologists, it is ensured that urgent and important cases can be processed early and in time; on the other hand, it can also optimize the workload distribution of radiologists, balance the workload distribution among radiologists, and assign competent radiologists to each medical image, so as to avoid delays in the processing of medical image-related data due to improper allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will be described hereinafter with reference to the accompanying drawings, in which like reference numerals refer to like elements, and in which:

[0077] Figure 1 A schematic diagram of a flow chart of a medical image scheduling method implemented by a computer according to an embodiment of the present invention;

[0078] Figure 2 FIG. 4 is a schematic diagram of the structure of a medical image scheduling system implemented by a computer according to an embodiment of the present invention.

[0079] Figure 3This is a user interface for radiologists according to another embodiment of the present invention, wherein medical image data are sorted and presented according to priority using a medical image scheduling method implemented by a computer according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0081] Various aspects and features of the present invention are described herein with reference to the accompanying drawings.

[0082] These and other characteristics of the invention will become apparent from the following description of a preferred form of embodiment given as a non-limiting example, with reference to the accompanying drawings.

[0083] It should also be understood that, although the invention has been described with reference to some specific examples, those skilled in the art will be able to realize many other equivalent forms of the invention that have the characteristics recited in the claims and are therefore within the scope of protection defined thereby.

[0084] The above and other aspects, features and advantages of the present invention will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.

[0085] Specific embodiments of the present invention are described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments of the invention are merely examples of the present invention, which may be implemented in a variety of ways. Well-known and / or repeated functions and structures are not described in detail to determine the true intent based on the user's historical operations and to avoid unnecessary or redundant details that obscure the present invention. Therefore, the specific structural and functional details invented herein are not intended to be limiting, but are merely used as the basis and representative basis for the claims to teach those skilled in the art to use the present invention in a variety of ways with substantially any suitable detailed structure.

[0086] This specification may use the phrases "in one embodiment", "in another embodiment", "in yet another embodiment" or "in other embodiments", all of which may refer to one or more of the same or different embodiments according to the present invention. The technical term "medical image-related data" used herein includes but is not limited to medical image data, and may also include other data related thereto, such as attribute information of the medical image (acquisition time, upload time, data source, etc.), such as the type of examination to which the medical image belongs, and also such as attribute information of the subject of the medical image (such as the subject's age, height, weight, gender, medical history, etc.).

[0087] As mentioned above, the current traditional medical image scheduling system can only arrange the priority according to the existing time order. Or it can only be determined by the clinician based on the patient's manifestation. In addition to the possibility of misjudgment by the doctor, the patient who should be in an emergency situation (such as cerebral hemorrhage) may lose the best treatment opportunity due to the slow processing flow. In view of this situation, the medical image scheduling method implemented by the computer proposed by the present invention can further process the medical image related data after obtaining the medical image related data, and then analyze it to obtain the priority parameters of the medical image related data. This analysis process can use the artificial intelligence module to make a preliminary judgment to evaluate the patient's condition. The final sorting can be carried out based on the priority parameters. Of course, the judgment of the clinician, the attribute information of the medical image including the data acquisition time, and the judgment of the artificial intelligence module can be further considered to sort the cases. In addition, for non-urgent cases, the system will also comprehensively consider the workload of the doctor during the scheduling process. In the following description, the priority score is used as an intuitive example of the priority parameter, but it should be noted that the priority parameter is not limited to this, but can be used in various ways, such as priority level, priority level, priority number, etc.

[0088] Specifically, Figure 1 As shown, in one embodiment of the present invention, a medical image scheduling method 100 implemented by a computer is provided, comprising the following steps:

[0089] Acquire medical image related data (step 101);

[0090] Using a processor, processing the medical image-related data to at least obtain a priority score of the medical image-related data (step 102);

[0091] sorting the medical image related data using a processor at least according to the priority scores of the medical image related data (step 103); and

[0092] Using a processor, a queue of medical image-related data and processing results thereof is presented according to the order (step 104 ).

[0093] The above-mentioned processor may be a processing device including one or more general-purpose processing devices such as a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), etc. More specifically, the processor may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The processor may also be one or more special-purpose processing devices such as an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a system on a chip (SoC), etc.

[0094] The processor may be communicatively coupled to the memory and configured to execute computer executable instructions stored therein, and each of the above steps may be stored in the form of computer executable instructions on the memory. The memory may include a read-only memory (ROM), a flash memory, a random access memory (RAM), a static memory, etc. In some embodiments, the memory may store computer executable instructions such as any one of steps 101-104 and data such as priority scores, the order of medical image-related data, and processing results thereof, which are used or generated when executing the computer executable instructions.

[0095] In the method provided in the embodiment of the present invention, the priority score involved is mainly used to evaluate the urgency of the patient's condition, that is, the priority of the patient's medical image data in the doctor's processing queue. Using the above method, the medical image-related data can be processed by a processor instead of manpower to obtain a priority score, and the medical image-related data can be sorted accordingly, and the real-time order of all medical image-related data to be processed can be presented to the radiologist. Different from relying on manpower, the processor can respond to the newly introduced medical image-related data in real time, and it can efficiently and accurately sort all the medical image-related data to be processed in near real time and display it to the radiologist, thereby avoiding the delay and error in the sorting relative to the actual situation caused by the lack or instability of manpower in the prior art.

[0096] Figure 2 A schematic diagram of the structure of the entire medical image scheduling system implemented by a computer according to an embodiment of the present invention is shown below. Figure 2The various modules in the system shown in the specific description of the various preferred embodiments of the computer-implemented medical image scheduling method described above. It should be noted that the description in combination with the various modules is only for the purpose of making the various steps of the preferred embodiments of the method easier to understand, and does not mean that the execution of the method must rely on such system and module settings. In fact, except for the image acquisition unit 110, each unit, module and platform are software units, software modules and software platforms defined by the functions performed. Therefore, according to the needs, the physical definition of the corresponding units, modules and platforms can be changed according to the functions performed, and is not limited to Figure 2 An example of the structure shown.

[0097] like Figure 2 As shown, the medical image scheduling system 100 includes: an image acquisition unit 110, which is used to acquire medical image related data and upload it to a data storage scheduling unit 130; the data storage scheduling unit 130, which is communicatively connected to the image acquisition unit 110 and the data processing unit 120, and performs data transmission, storage, sorting (based on the analysis results of the data processing unit 120), scheduling and recording results; the data processing unit 120, which receives the medical image related data transmitted from the data storage scheduling unit 130, processes and analyzes the data, and feeds back the analysis results to the data storage scheduling unit 130 for corresponding sorting and scheduling; and a data browsing unit 140, which is communicatively connected to the data storage scheduling unit 130, and provides various browsing interfaces of various display contents to possible users, including general radiologists, imaging department managers and developers, etc., according to the sorting and scheduling performed by it, and the display contents include queue presentation of medical image related data. Note that the image acquisition unit 110 is not necessarily included in the entire system 100 , and the system 100 may be communicatively connected to the image acquisition unit 110 to acquire medical image-related data therefrom and upload the data storage scheduling unit 130 .

[0098] The following is a detailed description of the steps performed in each section.

[0099] 1. Data storage scheduling unit 130

[0100] Preferably, the data storage scheduling unit 130 is responsible for receiving, transmitting, storing, sorting and scheduling medical image related data.

[0101] Receiving, transmitting and storing

[0102] Medical image-related data can be directly uploaded from the image acquisition unit 110 to the data storage scheduling unit 130, which is a local or remote imaging workbench. The acquired medical images include, but are not limited to, magnetic resonance imaging (MRI) images, 3D MRI, 2D fluidized MRI, 4D volumetric MRI, computed tomography (CT) images, cone beam CT, positron emission tomography (PET) images, functional MRI images (e.g., fMRI, DCE-MRI and diffusion MRI), X-ray images, fluorescence images, ultrasound images, radiotherapy portal images, single photon emission computed tomography (SPECT) images, etc.). For example, the image acquisition unit 110 can be an MRI imaging device, a CT imaging device, a PET imaging device, an ultrasound device, a fluoroscope device, a SPECT imaging device, or other medical imaging devices for obtaining one or more medical images of a patient. The acquired images can be stored in a database as imaging data and / or medical image-related data, and the data storage scheduling unit 130 can be communicatively connected to the database to retrieve the medical image-related data stored therein. As an example, the acquired images may also be stored in a memory included in the medical image scheduling system 100 , so as to be retrieved and used by various software modules including the data storage scheduling unit 130 .

[0103] In addition to the image acquisition unit 110, the data storage scheduling unit 130 can also obtain other medical image-related data from other units or systems. For example, the examination type to which the medical image acquisition belongs, the age, height, weight, gender, medical history, etc. of the examinee can be obtained from the hospital management information (HIS) system and the electronic medical record system.

[0104] The receiving of medical image-related data can be designed as a service based on a B / S (Browser / Server, i.e., a browser and server architecture, where the user interface is mainly implemented through a browser) structure. Users only need to log in with their own username and password to upload data. After the data is uploaded, the service will add a label to each data, including but not limited to the data acquisition time, upload time, data source, etc. After the label is added, the data enters the waiting queue and is ready to be fed into the data processing unit 120 in sequence by the data storage scheduling unit 130 to perform analysis and calculation and obtain analysis results.

[0105] According to the analysis results fed back by the data processing unit 120, the data storage scheduling unit 130 performs corresponding data transmission and scheduling. For example, when the feedback result is that the image quality does not meet the standard, the data storage scheduling unit 130 will forward the feedback result to the image acquisition unit 110. For example, for processable data, the data storage scheduling unit 130 will receive the automatic diagnosis results from the data processing unit 120. According to the sorting and scheduling rules, if there is a problem with the automatic diagnosis result, the data storage scheduling unit 130 will send the data (which may include the automatic diagnosis result) to the first doctor and wait for the result of the first doctor. After receiving the result of the first doctor, the automatic diagnosis result can be compared with the diagnosis result of the first doctor. If the results are inconsistent, the data storage scheduling unit 130 will hand over the result to another / or multiple doctors for comprehensive diagnosis to reduce the possibility of misdiagnosis. Note that the determination of whether there is a problem with the diagnosis result can be performed by the comprehensive analysis module 124 in the data processing unit 120 described below. The comprehensive analysis module 124 can instruct the data storage scheduling unit 130 to perform corresponding data transmission based on the determination result, such as sending the data (which may include automatic diagnosis results) to the first doctor and receiving the diagnosis results from the first doctor. The comprehensive analysis module 124 can compare the automatic diagnosis results with the diagnosis results of the first doctor, and instruct the data storage scheduling unit 130 to perform corresponding data transmission based on the comparison result, such as handing the result to another / or multiple doctors for comprehensive diagnosis, and then comprehensively analyzing the diagnosis results from each doctor to obtain the final data processing result, including the diagnosis result, priority score, etc.

[0106] The data storage scheduling unit 130 will return the final diagnosis result to the doctors and / or patients of the relevant hospital departments. Preferably, for incorrect diagnosis results, the data storage scheduling unit 130 will record them for reference in subsequent development.

[0107] Preferably, when the automatic diagnosis result is inconsistent with the final diagnosis result, the data storage scheduling unit 130 transmits the final diagnosis result data to the data processing unit 120. The data processing unit 120 can use the final diagnosis result data as training data to train the artificial intelligence analysis module therein to correct and improve it, thereby improving the accuracy of the analysis of the artificial intelligence analysis module.

[0108] The medical image scheduling system 100 can dynamically start / shut down computing nodes according to the number of cases (medical image-related data) in the waiting queue, so as to save computing resources and ensure computing efficiency. For example, when there are many cases in the waiting queue, the system 100 can start new nodes to improve computing efficiency; when there are no cases to be processed, the system 100 can shut down idle nodes to save resources.

[0109] Sorting Management

[0110] The data storage scheduling unit 130 also performs sequence management operations, and usually uses a priority list to dynamically sort the priority of medical image-related data. The list is sorted based on the priority score of the data. Generally speaking, the score can be set to a score with a fixed range (such as 0 to 100 points). The higher the score, the higher the sorting order in the queue. For new data entering the list, it is transmitted to the data processing unit 120 via the data storage scheduling unit 130, and the data processing unit 120 will calculate the priority score of the returned image-related data. The medical image-related data can be sorted directly according to the priority score calculated by the processor returned by the data processing unit 120 through image processing. In addition, other priority scores obtained based on other methods can also be used to sort the medical image-related data. For example, manually set scoring rules can also be accepted to generate a processing priority score. Table 1 gives a specific example of a scoring rule (a total score of more than 100 points is set to 100 points):

[0111] Table 1 Examples of scoring rules

[0112]

[0113] As shown in Table 1, the scores are different depending on whether the examination type is an emergency, etiology screening or a routine physical examination; different bonus items correspond to different bonus points. Table 1 does not list the scoring rules exhaustively, and the scoring rules may also consider other items. Among the priority scores obtained by the data processing unit 120 through image processing and the priority scores scored according to the manually predetermined rules, the higher score can be taken as the queuing score. Note that the step of scoring according to the manually predetermined rules may not be performed in the data storage scheduling unit 130, but may also be performed in the data processing unit 120 and the higher score may be taken as the final priority score and transmitted to the data storage scheduling unit 130.

[0114] In actual operation, the medical image-related data that cannot be processed by the data processing unit 120 can be directly set to be scored using predetermined rules, which can ensure that each medical image-related data is properly and timely processed in the queue.

[0115] In the actual queue scheduling process, the medical image-related data will be inserted into the corresponding position in the list according to the priority score. In another preferred embodiment, after the medical image-related data enters the queue, the priority score can be set to increase over time. This can ensure that no data is always in a queue-jumping state, that is, always in a low priority state, which can also reduce the additional risks to patients due to the long processing time to a certain extent. Of course, in this specific application, the growth rate of the priority score can depend on the initial score of the medical image-related data and the data processing speed range specified by the hospital. Such a design can also ensure that all data can be processed within the specified time limit. Specifically, for example, if the specified time limit is 120 minutes, the initial priority score of the data is 40 points, the full score is 100 points, and the time to process the data is less than 10 minutes, then the score of the data can increase by (100-40) / (120-10) / 60=0.009 per second. In order to ensure that urgent cases will always be at the top of the queue, when the score increases close to the full score, the growth rate will become a very small value to ensure that the score of non-urgent cases will not exceed 100 points. For example, when the score increases to 99, the growth rate will be reduced to 1e-5. For urgent cases, the initial score is 100, and the growth rate is also 1e-5. Maintaining a small growth rate is to ensure that the relative position of the data in the queue remains stable.

[0116] Scheduling Matching

[0117] The purpose of the above-mentioned queue adjustment based on priority scores is to enable doctors to see and process the medical image-related data of patients with more urgent conditions with higher priority. This process is actually related to the number of doctors who actually handle this queue.

[0118] Specifically, for smaller hospitals or medical image processing departments, when implementing the above method of the present invention, the complete queue of medical image-related data can be set to be visible to all doctors, so that each doctor can select data from the queue for processing according to the situation.

[0119] However, such a system would be difficult to manage for a department with a large number of people. For example, some data may remain in the queue without being processed, delaying the time for treating patients. For this situation, the system can be set to distribute data to individuals. The system will automatically distribute data to different doctors, each of whom has his own queue of data to be processed and is responsible for completing the data processing in a timely manner. The data storage scheduling unit 130 is responsible for matching data with doctors and transmitting specific data to the most appropriate radiologist for subsequent processing.

[0120] For example, the data can be matched with doctors through the following steps.

[0121] For example, medical image-related data is allocated according to the doctor's current work progress. Urgent data (with a higher priority score) will be allocated to the doctor with the least urgent cases on hand; ordinary data will be allocated to the doctor who is most proficient in it for processing. If the doctor who is proficient in this type of data has too much workload and cannot accept more data, the data will be allocated to the doctor with the lightest task to complete. Preferably, the data storage scheduling unit 130 can also monitor the processing of the data. When the specified deadline of the data is approaching and the assigned doctor has not yet processed it, the data storage scheduling unit 130 will provide a reminder to the assigned doctor in a visual form such as highlighting. Preferably, in this case, the data storage scheduling unit 130 will present the situation that the assigned doctor has not yet processed the data with an approaching deadline on the interface of the imaging department manager. Preferably, in this case, the data storage scheduling unit 130 can automatically assign the task to another imaging doctor at the current priority level to ensure that the processing results can be obtained on time.

[0122] 2. Data processing unit 120

[0123] The medical image related data is transmitted to the data processing unit 120 via the data storage scheduling unit 130, wherein the medical image related data is processed by a processor to at least obtain a priority score of the medical image related data. The data processing unit 120 may include an image analysis module 123 for processing the medical image related data to at least obtain a priority score of the medical image related data. Preferably, Figure 2 As shown, the data processing unit 120 may include an image filter 121 to filter the medical image-related data before processing it, so as to ensure that different categories of data can be properly and efficiently processed respectively, and to avoid wasting computing resources and time on unprocessable data, which may delay the treatment process. More preferably, the data processing unit 120 may also include an image sorter 122, which is used to classify the medical image-related data according to the modality and the imaging part, so as to select a suitable image analyzer in the image analysis module 123 to process the data, so as to at least obtain the priority score of the medical image-related data. More preferably, the data processing unit 120 may also include a comprehensive analysis module 124, which comprehensively analyzes the processing results of the image analysis module 123 with other information, so as to obtain a more reliable final priority score and reduce the possibility of misdiagnosis.

[0124] Figure 2The data processing unit 120 is shown to include an image filter 121, an image sorter 122, an image analysis module 123, and a comprehensive analysis module 124. It should be noted that the architecture of the data processing unit 120 is not limited thereto, and the image filter 121, the image sorter 123, and the comprehensive analysis module 124 are all preferred but not essential modules. The steps performed by each unit are described in detail below.

[0125] Image Filter 121

[0126] Preferably, the acquired medical image-related data is first filtered by the image filter 121, and the filtered processable data is subsequently processed by the processor (eg, the image sorter 122, the image analysis module 123, and the comprehensive analysis module 124).

[0127] Specifically, the acquired medical image-related data is screened by the image filter 121, and the data is usually divided into several categories: invalid data, processable data, data of unqualified quality, and unknown data. Invalid data refers to data where the image does not exist or the data is incomplete, and data of unqualified quality refers to data such as blank or data with too low signal-to-noise ratio. When encountering such data, the data processing unit 120 will immediately notify the user so that the user can process it in time. If no updated data is transmitted back from the image acquisition unit in a short period of time, the data will be put into a list according to the rules established in advance and wait for the radiologist to confirm and process it. Unknown data refers to data of a type that has never appeared in the training data set of the classifier in the data processing unit 120 (for example, it may be used for the image sorter 122, or it may also be used for the image analysis module 123). For example, if the training data set of the classifier is various types of human CT images, then for the classifier, images of other modalities or medical images of animals will be regarded as unknown data. Unknown data will be put into the list of data to be processed according to the rules established in advance and wait for the radiologist to process it. At the same time, the data will also be recorded for reference for subsequent development. The processable data will be directly transferred to the subsequent automatic data processing units, such as the image sorter 122, the image analysis module 123 and the comprehensive analysis module 124, for subsequent processing.

[0128] The image filter 121 may be implemented based on a variety of different algorithms.

[0129] For example, the signal-to-noise ratio of an image can be calculated by comparing the original image and the smoothed image to exclude images with excessive noise.

[0130] For example, the entropy value of the image can be calculated (smoothed) to exclude noise images (such as blank images) that do not contain valid information.

[0131] For another example, the distribution of training data can be obtained by applying a one-class support vector machine to the training data set. A one-class support vector machine is an unsupervised learning algorithm based on the support vector machine algorithm. The algorithm learns all of our training data and calculates the characteristic distribution interval of the data. For the data to be processed, the trained one-class support vector machine determines whether the new data conforms to the distribution of the existing data to determine whether the data is processable data or unknown data.

[0132] Image sorter 122

[0133] Preferably, the data marked as processable by the image filter 121 will be transmitted to the image sorter 122 for automatic sorting. The image sorter mentioned in the present invention is a module 122 for realizing the classification of the modality and imaging organ of the medical image-related data according to the characteristics of the image. In fact, due to different modalities and different imaging parts, the image data itself has regular differences. An image analyzer suitable for a certain modality (such as CT) and a certain imaging part (such as breast) may not be suitable for image analysis of other modalities (such as MR) and other imaging parts (such as the brain). In view of this problem, preferably, multiple image analyzers can actually be provided, corresponding to the modalities and imaging parts of the medical image respectively. Therefore, in this process, it is necessary to select the most appropriate image analyzer to analyze the image.

[0134] In addition, in the traditional medical image analysis process, the image classification mentioned above is often judged by the meta information generated by the image acquisition device and the text information input by the acquisition personnel. Since different equipment manufacturers will adopt different meta information generation standards, different acquisition personnel will have different text input habits, and different types of data will be mixed together (such as brain CT images and chest CT images will be mixed together), this type of pre-classification method based on non-image information is not reliable. According to the image sorter 122 involved in the method in the embodiment provided by the present invention, when it is specifically executed, the modality of the image and the imaged organ will be judged based on the image features. At the same time, the image sorter 122 designed by us can divide the image into different sub-images according to the organ part and hand them over to different image analyzers for processing. For example, chest CT images often contain 3D images of different organs such as lungs, spine, ribs, heart, etc. At this time, as a further preferred embodiment, the original image can be first segmented to obtain image blocks of different organs, and then the image blocks of each organ can be handed over to the corresponding different image analyzers for targeted analysis.

[0135] The image sorter 122 may be implemented by different computer program algorithms, including but not limited to image retrieval algorithms and image classification algorithms.

[0136] For example, the image sorter 122 is specifically implemented by an image retrieval algorithm. The basic process of this algorithm is: a set of feature values ​​is generated for each sample image and the image to be processed, and the closest sample image is found by matching the feature values ​​of the image to be processed and the feature values ​​of the sample image to achieve the purpose of classification. Further, the extraction of the above-mentioned image feature values ​​can use traditional algorithms such as the histogram of directional gradients (HOG) feature algorithm, the local binary (LBP) feature algorithm, and the Haar feature algorithm, or a deep convolutional network based on multi-layer convolution. Specifically, the deep convolutional network can be obtained by supervised classification task training or unsupervised automatic encoder training. Preferably, in order to effectively reduce the feature dimension to improve the matching efficiency, a feature cooling algorithm can be used. Traditional algorithms include principal component analysis (PCA), independent component analysis (ICA), dictionary learning combined with sparse encryption algorithm (dictionary learning and sparse coding), etc., and more advanced machine learning algorithms include bag-of-words models, word vector models (such as word2vec), etc.

[0137] As an alternative, the image sorter designed in the embodiment of the present invention can also be implemented using an image classification algorithm. For example, images can be classified by training a deep convolutional network to classify images into different organs and different modalities. The input of this type of deep convolutional network is the original image, and the output is the probability value that the image belongs to each specified type. The deep convolutional network first extracts image feature information through a combination of multiple convolutional layers, and then calculates the final probability value through multiple fully connected layers. The convolution kernel and connection weights in the network are obtained by computer optimization. The deep convolutional network can also predict the existence of multiple targets in an image at the same time and find the corresponding sub-areas. Different sub-graphs can be generated for each sub-area to further sort the original image. Note that this type of deep convolutional network can also be implemented as a full convolutional network, using the high-speed computing characteristics of convolution to speed up the computing speed.

[0138] Image analysis module 123

[0139] Preferably, the image analysis module 123 involved in the method of the embodiment of the present invention is composed of multiple image analyzers. Each image analyzer is designed to perform calculations for a specific modality and a specific organ. The image analyzer can be an artificial intelligence algorithm based on deep learning (neural network), or other artificial intelligence algorithms, such as one or more algorithms in deep neural network, random forest, gradient boosting decision tree. The result of the image analyzer processing the medical image related data includes the diagnosis result, relevant area, and priority score of the medical image. The diagnosis result refers to the measurement of relevant pathological analysis and physiological indicators based on the image. For example, the diagnosis result can be the probability that the image feature shows a certain disease feature, or it can be the size, volume, morphological analysis, and tissue structure analysis of a specific organ. The relevant area refers to the area of ​​interest. The relevant area can be annotated by one or more images. Each pixel in the image can be a number value for different areas or a probability value that the pixel belongs to a certain area. The relevant area can also be an area surrounded by one or more computer graphics (such as straight lines, curves, boxes, circles). The priority score reflects the time urgency of the subject's condition. For patients with urgent conditions that require immediate assistance (such as cerebral hemorrhage), the image analyzer will generate a high priority score, while images with no obvious symptoms will be given a lower priority score.

[0140] As an example, the diagnosis result in the processing result can be obtained by using an image classification algorithm similar to the image filter 121. When a classification calculation based on a deep convolutional network is used, the network inverse algorithm can also be used to reversely infer the area in the image that brings about the decision, thereby locating the potential area of ​​the lesion. The network inverse algorithm reverses the descending gradient value used for network parameter optimization to the variable value transmitted in the network, thereby inferring the area that has a positive impact on a certain decision (such as a potential lesion area).

[0141] As an example, the relevant areas in the processing results can be obtained by using an image segmentation algorithm, for example, by detecting lesions and segmenting relevant areas through an image segmentation algorithm. This algorithm can be a deep learning algorithm using a fully convolutional network. The input of the algorithm is the original image, and the output is a probability distribution image of the segmented object. In the middle, multiple convolution layers are used to extract relevant image features so that the final judgment can be made based on the image features. This algorithm can not only be used to segment lesion areas, but also to segment areas of human organs in images.

[0142] As an example, the priority score in the processing result can be obtained based on the diagnostic condition of the medical image (such as the diagnostic result obtained by the image analyzer mentioned above, or the diagnostic classification shown in Table 1) according to pre-established rules. Specifically, for each condition, such as the diagnosis result of 80% probability of intracerebral vascular obstruction, or the emergency examination of brain images (such as emergency examination), there will be a corresponding score. As an alternative, a regression algorithm can also be used to directly calculate the priority score from the image features. The regression algorithm can use an algorithm based on a deep convolutional network, which replaces the original training output with a priority score. It can also be calculated using an algorithm such as a linear regression tree, which uses the image features obtained by convolution as input, and finally obtains the corresponding score by combining the features. Compared with using established rules, the linear regression algorithm is more flexible.

[0143] The scoring rules for priority scores can be set by doctors. For example, for emergency cases (immediate treatment after admission), the priority score is 100 points. For patients with healthy examination results, the priority score is 0. For patients with chronic diseases, the severity of the disease can be determined by whether surgery is required, medication status, and rehabilitation effect, and then scored.

[0144] Comprehensive Analysis Module 124

[0145] As a further preference, the analysis results of the image analysis module 123 can be further comprehensively analyzed, which can be achieved by the comprehensive analysis module 124. Specifically, this process comprehensively analyzes the diagnostic results of the medical image-related data, the image features of the medical image, and the attribute information of the patient of the medical image. If there are different data for the same patient, the comprehensive analysis module will combine the different data for the final diagnosis. For example, if the same patient has medical image-related data of different modalities and different parts, the processing results of these data can be combined to make a final diagnosis. For example, if the ultrasound image shows that there is a divergent and active blood flow in a certain place, and the CT image shows that there is a heterogeneous structure in the same place, then the combined diagnosis can be obtained that there is a high probability of a tumor in that place. For another example, if the CT image and ultrasound image of other organs of the same patient also find multiple heterogeneous structures with active blood flow, then the combined diagnosis can be obtained that there is tumor metastasis, and accordingly, the priority score is higher.

[0146] On the other hand, the comprehensive analysis module 124 can also integrate other information to reduce the possibility of misdiagnosis. For example, the image analyzer determines that a patient has a risk of Alzheimer's disease by processing the brain MRI image, but according to the case information, it is found that the patient is only 18 years old. In this case, the diagnosis result is not reliable and needs to be further revised. This medical record information can come from medical image related data.

[0147] Preferably, the comprehensive analysis module 124 can perform comprehensive analysis on various priority scores of the same medical image-related data obtained in different ways, such as priority scores calculated for image features by various artificial intelligence algorithms, priority scores determined for attribute information of medical images according to artificially defined rules, priority scores manually diagnosed by different doctors, and priority scores determined according to predetermined rules based on the diagnosis results of medical images, etc. The comprehensive analysis may include any one or more of discarding deviating values, taking the highest value, taking the average value, etc.

[0148] The above-mentioned medical image-based diagnostic results can be integrated with other information through Bayesian networks, hidden Markov models, and random forest algorithms to obtain the final probability value of the disease. These algorithms combine the image-based diagnostic results and the image features used to calculate the diagnostic results with other non-image information (such as age, height, weight, gender, etc.) as the characteristic values ​​of the calculation input. The output is the judgment result or probability value that the patient suffers from a certain (or multiple) disease. Bayesian networks and hidden Markov models calculate the corresponding probability model through training data to obtain the final output probability value. The random forest algorithm generates multiple decision trees by randomly sampling the characteristic values ​​and samples for voting to generate the final judgment result.

[0149] As a further preference, in the scheduling method provided by the embodiment of the present invention, the priority score is the main reference for dynamically adjusting the queue order. In the specific implementation, the medical image-related data can be mainly processed by means based on artificial intelligence to obtain the priority score, diagnosis result and related area of ​​the medical image-related data. However, if necessary, a correction mechanism can also be introduced, that is, when the comprehensive analysis module 124 finds that there is a problem with the diagnosis result through comprehensive analysis, the obtained diagnosis result can be corrected in various ways. Specifically, the correction includes: sending the medical image-related data to the first doctor and obtaining the diagnosis result of the first doctor; comparing the diagnosis result obtained by the image analysis module 123 with the diagnosis result of the first doctor; and obtaining the diagnosis result, related area and priority score of the medical image-related data based on the comparison result. In this process, the step of obtaining the diagnosis result, related area and priority score of the medical image-related data based on the comparison result specifically includes: when the comparison results are inconsistent, sending the diagnosis results of the two to other doctors; and obtaining the diagnosis result, related area and priority score of the medical image-related data based on the comprehensive diagnosis result of the doctor.

[0150] 3. Data browsing section 140

[0151] According to the sorting completed by the data storage scheduling unit 130, a queue of medical image-related data and its processing results is presented on the data browsing unit 140. Figure 3 As shown, the data browsing unit 140 may provide multiple user interfaces to provide corresponding displays for different users.

[0152] Radiologist User Interface141

[0153] Preferably, a series of sorted patient image thumbnails are displayed in the radiologist user interface 141 . Figure 3 FIG. 1 shows a user interface 141 for radiologists according to another embodiment of the present invention. Figure 3 As shown, radiologists can classify and display pending cases through various classification modes, including but not limited to modality classification, organ classification, age classification, task difficulty, etc. Each radiologist has different preferences. Some are good at diagnosing a certain part or organ, such as the lungs, so they tend to classify pending cases according to organs. Some are selective about imaging modalities. For example, when radiologists who are not good at CT image processing find that they have a large number of urgent CT images through modality classification, they can feedback or submit them to higher-level doctors through the user interface 141 for redistribution or processing, so as to achieve autonomous optimization of human resources configuration.

[0154] For example, if you select modal classification, Figure 3 As shown, the cases to be processed are divided into various modalities such as CT images, MRI images, and X-ray images for display. In each modality, the medical image-related data are sorted from left to right according to the priority scores from high to low. For example, in the CT image column, the leftmost is the "Urgent!!!" case, and the cases to the right are higher priority cases. A dynamic priority score is displayed below each medical image. Preferably, a bar is displayed near the priority item, and the length of the bar represents the size of the priority score, so that the radiologist can more intuitively grasp the priority of the cases to be processed. Preferably, the radiologist opens the corresponding case for diagnosis by clicking on the thumbnail. The thumbnail is generated by the data processing unit 120 based on the patient's medical image to be diagnosed. For example, the data processing unit 120 can select the optimal contrast, key layers, etc. to display the medical image. Preferably, the data processing unit 120 will adjust the image brightness and select the displayed area according to the calculation processing results. If a certain disease is detected, the relevant area will be highlighted. As shown Figure 3As shown in the X-ray image column, the data processing unit 120 generates highlight annotations for the relevant areas in the lungs to provide diagnostic assistance to radiologists. Preferably, when the mouse hovers over an image, the image will be dynamically displayed and automatically switched to different disease images. Preferably, radiologists can choose to display the calculation results of the data processing system to narrow the scope, improve work efficiency, and reduce the misdiagnosis rate. If the doctor encounters difficulties in processing the data, it can be submitted to a higher-level doctor for processing through the user interface 141. The data will then enter the queue of data to be processed by the higher-level doctor with the current priority score.

[0155] Radiology Manager User Interface 142

[0156] The imaging department manager user interface 142 allows the department manager to monitor and schedule data progress through the user interface. The manager can view the data to be processed, the data processed, and the statistics of the processed data of each imaging doctor through the user interface. The manager can call out the processed data for random inspection, and can also manually allocate the data to be processed by different doctors.

[0157] Results Browsing User Interface 143

[0158] The result browsing user interface allows non-radiologists to view the diagnosis results. At the same time, within the scope permitted by hospital regulations, patients can also use the result browsing user interface to conveniently access and view their own imaging data, data processing progress, and data processing result reports, such as on their own mobile terminals.

[0159] Developer User Interface 144

[0160] Preferably, developers can view the error diagnosis result statistics of the artificial intelligence module through the developer user interface 144, so as to provide reference for system error correction and subsequent system upgrades. In addition, an embodiment of the present invention also provides a computer-readable storage medium storing computer executable instructions, which implement any example of the above method when executed by a processor. The basic embodiment of the method includes:

[0161] Obtain medical image-related data;

[0162] Processing the medical image-related data using a processor to at least obtain a priority score of the medical image-related data;

[0163] sorting, using a processor, the medical image related data based at least on a priority score of the medical image related data; and

[0164] A processor is used to present a queue of medical image-related data and processing results thereof according to the sorting.

[0165] Various specific embodiments of the method described above, including various software modules, can be implemented on the computer-readable storage medium.

[0166] Above, various operations or functions are described herein, which can be implemented or defined as software codes or instructions. Such content can be directly executable ("object" or "executable" form) source code or differential code ("increment" or "patch" code). The software implementation of the embodiments described herein can be provided via a product in which code or instructions are stored or via a method of operating a communication interface to send data via a communication interface. A machine or computer readable storage medium can enable a machine to perform the described functions or operations, and includes any mechanism for storing information in a form accessible to a machine (e.g., a computing device, an electronic system, etc.), such as a recordable / unrecordable medium (e.g., a read-only memory (ROM), a random access memory (RAM), a disk storage medium, an optical storage medium, a flash memory device, etc.). The communication interface includes any mechanism for connecting to any of the hardwired, wireless, optical, etc. media to communicate with another device, such as a memory bus interface, a processor bus interface, an Internet connection, a disk controller, etc. The communication interface can be configured to prepare the communication interface to provide a data signal describing the software content by providing configuration parameters and / or sending a signal. The communication interface may be accessed via one or more commands or signals sent to the communication interface.

[0167] The present invention also relates to a system for performing the operations described herein. The system may be specially constructed for the desired purpose, or the system may include a general-purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer-readable storage medium, such as but not limited to any type of disk including a floppy disk, an optical disk, a CDROM, a magneto-optical disk, a read-only memory (ROM), a random access memory (RAM), an EPROM, an EEPROM, a magnetic card or an optical card, or any type of medium suitable for storing electronic instructions, each of which is coupled to a computer system bus.

[0168] The above embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention. The protection scope of the present invention is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the essence and protection scope of the present invention, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present invention.

Claims

1. A medical image scheduling method, which is implemented by a computer, and It is characterized in that The medical image scheduling method comprises the following steps: Obtain medical image-related data; Using a processor, processing the medical image-related data to obtain at least a priority parameter of the medical image-related data, wherein the priority parameter is used to assess the urgency of the patient's condition; dynamically sorting the medical image related data using the processor at least according to a priority parameter of the medical image related data; Using the processor, presenting a queue of medical image-related data and processing results thereof according to the sorting; and Using the processor, matching the medical image-related data with doctors, this process is performed in association with the doctor's current work progress and / or the urgency of the condition, specifically comprising: allocating the medical image-related data to each doctor according to the priority parameter of the medical image-related data and the current work progress of each doctor, wherein the medical image-related data with a priority parameter corresponding to an emergency case is allocated to the doctor with the least emergency cases in the queue of data to be processed; The processor is used to monitor the processing status of the medical image related data; and when it is monitored that the prescribed deadline of the medical image related data is approaching and the assigned doctor has not yet processed it, a reminder is provided to the assigned doctor and the medical image related data that is approaching the prescribed deadline and has not been processed is automatically assigned to other doctors with its current priority parameter; wherein, The priority parameter is expressed as a priority score and increases over time, the growth rate of the priority score depends on the initial score of the medical image-related data, the full score and the data processing speed range specified by the hospital, and, When the priority score of a non-urgent case grows close to the full score, the growth rate of its priority score is changed to a very small value to ensure that the priority score of the non-urgent case does not exceed the initial score of the priority score of the urgent case.

2. The medical image scheduling method according to claim 1, It is characterized in that Further including: The process of matching the medical image related data with doctors is performed in association with the number of doctors processing the queue.

3. The medical image scheduling method according to claim 2, It is characterized in that The performing in association with the number of doctors processing the queue further comprises: For smaller hospitals or medical image processing departments, the queue of medical image-related data is set to be visible to all doctors; For departments with a large number of people, the medical image-related data are automatically distributed to different doctors, so that each doctor has his own queue of data to be processed.

4. The medical image scheduling method according to claim 1, It is characterized in that According to the priority parameters of the medical image related data and the current work progress of each doctor, allocating the medical image related data to each doctor further includes: Assign medical image-related data for non-urgent cases to specialized physicians; The priority parameter of the non-emergency case is lower than the priority parameter of the emergency case.

5. The medical image scheduling method according to claim 4, It is characterized in that Also included is that the processor: In the case that the doctor who is good at handling the non-emergency case cannot receive more medical image-related data for processing, the medical image-related data of the non-emergency case is allocated to the doctor with the least current workload.

6. The medical image scheduling method according to claim 1, It is characterized in that The reminder is displayed in a highlighted visual form.

7. The medical image scheduling method according to claim 1, It is characterized in that It also includes that the processor: classifies and displays the cases to be processed in the queue of data to be processed according to the classification mode.

8. The medical image scheduling method according to claim 7, It is characterized in that The classification mode includes at least one of modality classification, organ classification, age classification, and task difficulty.

9. A medical image scheduling system, the medical image scheduling system being communicatively connected to an image acquisition unit to obtain medical image related data therefrom, It is characterized in that The medical image scheduling system comprises: a data processing unit configured to process the acquired medical image-related data to obtain at least a priority parameter of the medical image-related data, wherein the priority parameter is used to evaluate the urgency of the patient's condition; a data storage scheduling unit configured to dynamically sort the medical image related data according to at least a priority parameter of the medical image related data; and A data browsing section configured to present a queue of medical image-related data and processing results thereof according to a sorting order; The data storage scheduling unit is further configured to: match the medical image-related data with doctors, and this process is performed in association with the current work progress and / or urgency of the doctor's condition, specifically including: assigning the medical image-related data to each doctor according to the priority parameter of the medical image-related data and the current work progress of each doctor, wherein the medical image-related data with a priority parameter corresponding to an emergency case is assigned to the doctor with the least emergency cases in the queue of data to be processed; monitoring the processing of the medical image-related data; and, when it is monitored that the prescribed deadline of the medical image-related data is approaching and the assigned doctor has not yet processed it, providing a reminder to the assigned doctor and automatically assigning the medical image-related data with a prescribed deadline approaching and not yet processed to other doctors with its current priority parameter; wherein the priority parameter is expressed as a priority score and grows over time, and the growth rate of the priority score depends on the initial score of the medical image-related data, the full score, and the data processing speed range specified by the hospital, and, when the priority score of a non-emergency case grows to a close to full score, the growth rate of its priority score is changed to a very small value to ensure that the priority score of the non-emergency case does not exceed the initial score of the priority score of the emergency case.

10. The medical image scheduling system according to claim 9, It is characterized in that The data storage scheduling section is further configured to perform a process of matching the medical image-related data with doctors in association with the number of doctors processing the queue.

11. The medical image scheduling system according to claim 10, It is characterized in that The data storage scheduling unit is further configured as: For smaller hospitals or medical image processing departments, the queue of medical image-related data is set to be visible to all doctors; For departments with a large number of people, the medical image-related data are automatically distributed to different doctors, so that each doctor has his own queue of data to be processed.

12. The medical image scheduling system according to claim 9, It is characterized in that The data browsing section comprises: A radiologist user interface, which is configured to classify and display the pending cases in the pending data queue of each radiologist according to the classification mode; A result browsing user interface configured for non-radiologists or patients to view the diagnosis results; and A developer user interface is configured for developers to view error diagnosis result statistics of the artificial intelligence module.

13. The medical image scheduling system according to claim 9, It is characterized in that The data browsing section also includes an administrator user interface, which is configured for the administrator to monitor and schedule data progress, wherein the monitoring and scheduling of data progress includes: checking the data to be processed by each radiologist, the processed data and the statistics of the processed data, calling out the processed data for random inspection, and manually allocating data to be processed by different doctors.

14. A computer-readable storage medium storing computer-executable instructions, wherein when the computer-executable instructions are executed by a processor, the computer-implemented medical image scheduling method according to any one of claims 1 to 8 is implemented.