Remote collaboration and expert guidance system for nuclear magnetic resonance
By introducing parameter presetting, image quality checking, intelligent image segmentation, real-time control, and expert relay mechanisms into the nuclear magnetic resonance system, the problems of low efficiency and resource waste in remote nuclear magnetic resonance collaboration have been solved, and efficient remote guidance and image processing have been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUNAN MAGTECH MEDICAL TECH CO LTD
- Filing Date
- 2026-04-29
- Publication Date
- 2026-05-29
AI Technical Summary
In existing technologies, remote collaboration and guidance methods for nuclear magnetic resonance imaging are inefficient, wasteful of expert resources, suffer from unstable image transmission, high communication costs due to language misunderstandings, and cannot efficiently utilize expert resources.
Through the parameter preset module, image quality module, data transmission module, remote adjustment module, remote guidance module, and expert relay module, efficient collaboration between remote experts and primary hospitals is achieved, including parameter review, image self-repair, intelligent image segmentation, real-time control signal processing, and expert relay mechanism.
It improves the efficiency of remote MRI collaboration and the utilization of expert resources, reduces wasted time on communication and guidance, ensures image quality, and optimizes the use of expert resources.
Smart Images
Figure CN122117311A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of medical communications, and in particular to a remote collaboration and expert guidance system for MRI. Background Technology
[0002] Magnetic resonance imaging (MRI) is one of the core tools for precision diagnosis in modern medicine. However, high-quality MRI scans heavily rely on the precise setting of complex scanning sequence parameters and real-time control of image quality during the scanning process.
[0003] Existing technologies typically utilize remote collaboration and guidance models. However, traditional solutions often employ remote desktop sharing and video calls, with experts from higher-level hospitals providing hands-on online guidance to junior technicians. On one hand, this approach wastes a significant amount of time on basic tasks and consumes substantial expert resources. Furthermore, image unavailability during image processing also wastes considerable medical resources. On the other hand, in conventional network environments, data transmission may be unstable, and misunderstandings of expert verbal guidance can arise due to language barriers or other influencing factors, leading to excessively high communication costs. Therefore, how to intelligently and efficiently utilize expert resources for effective remote collaboration in MRI scans has become a pressing issue. Summary of the Invention
[0004] The purpose of this invention is to provide a remote collaboration and expert guidance system for nuclear magnetic resonance imaging (MRI) to solve the problems mentioned in the background art.
[0005] This application provides a remote collaboration and expert guidance system for nuclear magnetic resonance imaging (MRI), the system comprising: Parameter preset module: used to acquire basic patient information, examination site and previous images, generate scan preset parameters, and remote experts review and adjust the scan preset parameters to obtain the target scan preset parameters; Image quality module: used to perform MRI examination on the patient based on the preset parameters of the target scan, and to perform image quality check during the image generation process to identify image problems, generate self-repair guidelines based on the image problems, and perform image processing according to the self-repair guidelines to obtain the target MRI image; Data transmission module: used to acquire the communication channel between the hospital and remote experts, obtain the channel bandwidth based on the communication channel, perform intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain the core image, and send the core image to the remote experts; Remote adjustment module: used to acquire voice data and gesture data of remote experts, recognize the voice data and gesture data to obtain real-time control signals and real-time guidance information from remote experts, and adjust the parameters of the MRI equipment according to the real-time control signals; Remote guidance module: Based on the real-time guidance information, it performs spatial positioning on the real screen of the hospital's operating terminal, generates visual guidance, and guides the operator to perform operations; Expert relay module: Used to establish a collaborative task queue, which contains multiple remote experts. When a remote expert leaves, the module packages the remote expert's annotation adjustment records, problem analysis records, and guidance records into a standardized information package and transfers it to another available expert in the collaborative task queue.
[0006] Preferably, the steps of acquiring basic patient information, examination sites, and previous images, generating preset scanning parameters, and having a remote expert review and adjust the preset scanning parameters to obtain the target preset scanning parameters are as follows: Obtain basic patient information, examination sites, and previous images; generate basic examination parameters based on the patient's basic information. Based on the examined area and the previous images, the patient's problem points are located to obtain the lesion location. Based on the lesion location, the scanning depth and scanning area are determined. By combining the basic inspection parameters, the scanning depth, and the scanning area, preset scanning parameters are generated; The preset scanning parameters are sent to a remote expert, who reviews and adjusts them to obtain the target preset scanning parameters before sending the data back.
[0007] Preferably, the steps of performing image quality checks during image generation to identify image problems, generating self-repair guidelines based on the image problems, and performing image processing according to the self-repair guidelines to obtain the target MRI image are as follows: The image generation process is monitored to obtain the image data stream before image generation, and the image data stream is quality identified to obtain data stream quality points and the problem data streams corresponding to the data stream quality points; Identify the image problem type, image problem location, and image problem phenomenon corresponding to the data stream quality point to obtain the image problem; Based on the image problem, the problem source is traced to obtain the data output process of the problem data stream, and the anomaly monitoring of the data output process is performed to obtain the target anomaly point; Extract abnormal data from the target anomaly point and generate a self-repair guide based on the abnormal data output operation; Based on the self-healing guidelines, an image repair operation is generated, and the image problem is processed according to the image repair operation to obtain the target MRI image.
[0008] Preferably, the steps of obtaining the channel bandwidth based on the communication channel, performing intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain a core image, and sending the core image to a remote expert are as follows: When the signal is first established between the hospital and the remote expert, the communication channel between the hospital and the remote expert is obtained, and the bandwidth of the communication channel is tested to obtain the channel bandwidth. Based on the channel bandwidth, the upper limit of the communication rate of the communication channel is obtained, and a redundancy amount is set according to the upper limit of the communication rate. The channel bandwidth is then set according to the redundancy amount to obtain the target channel bandwidth. The image data volume of the target NMR image is obtained, and the minimum channel bandwidth required for the target NMR image is obtained based on the image data volume; Determine whether the minimum channel bandwidth is greater than the target channel bandwidth; If it is determined that the minimum channel bandwidth is greater than the target channel bandwidth, then the target MRI image is identified to obtain key image information; Based on the key image information, the target MRI image is cropped to obtain a core image, which is then sent to a remote expert.
[0009] Preferably, the step of sending the core image to a remote expert specifically includes: During the transmission of the core image, the network quality of the communication channel is monitored to obtain real-time network quality parameters; Determine whether the real-time network quality parameter is consistently lower than the preset standard network quality parameter within a preset time period; If it is determined that the network quality parameter is consistently lower than the preset standard network quality parameter within a preset time period, then key data information in the core image is extracted. The key data information is packaged and compressed to generate a key data packet, which is then sent to a remote expert. Extract secondary data information that is not required in real time from the core image, in addition to the key data information, and resume transmission of the secondary data information from breakpoints.
[0010] Preferably, the steps of acquiring voice and gesture data from a remote expert, recognizing the voice and gesture data to obtain real-time control signals and real-time guidance information from the remote expert, and adjusting the parameters of the MRI equipment based on the real-time control signals are as follows: Acquire voice and gesture data from remote experts; Based on the voice data, voice information recognition is performed on the voice data to obtain text content, and keyword extraction and context recognition are performed on the text content to obtain language instruction information; Based on the gesture data, gesture information recognition is performed on the gesture data to obtain hand posture information, posture change information, and hand trajectory information; Based on the hand posture information, the posture change information, and the hand trajectory information, the gesture command information of the remote expert is obtained; By combining the language instruction information and the gesture instruction information, the instruction information is collected and divided to obtain real-time control signals and real-time guidance information. The real-time control signal is transmitted back to the hospital's MRI equipment to adjust the parameters of the MRI equipment.
[0011] Preferably, based on the real-time guidance information, spatial positioning is performed on the actual screen of the hospital's operating terminal to generate visual guidance and guide the operator through the operation steps, specifically as follows: The system acquires real-time images from the hospital's control panel, identifies these images, and obtains multiple control points and control areas. Based on the real-time guidance information, multiple operation points and operation areas are matched to obtain target operation points and target operation areas; Extract the actual spatial location of the target operation point and the target operation area, as well as the corresponding operation action; Based on the actual spatial location and the operation action, a visual guide is generated and projected onto an AR display device preset on the operation terminal.
[0012] Preferably, the process of establishing a collaborative task queue, in which multiple remote experts are selected, and when a remote expert leaves, packaging their annotation adjustment records, problem analysis records, and guidance records into a standardized information package, and transferring it to another available expert in the collaborative task queue, specifically involves: Establish a collaborative task queue, in which multiple remote candidate experts are set, and the current idle status of each remote candidate expert is marked; When a remote expert leaves, the system automatically retrieves the expert's annotation and adjustment records, problem analysis records, and guidance records, and generates a standardized information package. Based on the standardized information package, the current actual inspection progress is identified, and a progress docking interface is generated based on the actual inspection progress. In the collaborative task queue, a remote candidate expert whose current status is marked as idle is found in sequence and marked as the target remote expert; The standardized information package and the progress docking interface are sent to the target remote expert, and the standardized information package is extracted and simplified to generate a quick reading report.
[0013] In summary, this application includes at least one of the following beneficial technical effects: By acquiring the patient's basic information, examination site, and previous images, preset scanning parameters are generated and sent to remote experts for review and adjustment to obtain the target scan preset parameters. The patient then undergoes an MRI examination. During image generation, image quality is checked to identify image problems. Based on these problems, a self-repair guide is generated and data repair is performed to obtain the target MRI image. Next, the communication channel between the hospital and the remote expert is acquired, and its bandwidth is determined. Based on the channel bandwidth, the target MRI image is segmented and sent to the remote expert. Then, the remote expert's voice and gesture data are acquired to generate real-time control signals and guidance information. The MRI equipment parameters are adjusted based on the real-time control signals. Based on the real-time guidance information, spatial positioning is performed on the real-world screen at the hospital's operating terminal to generate visual guidance for the operator. Finally, a collaborative task queue is established with multiple remote experts as candidates. If the current remote expert needs to disconnect, a target remote expert is selected from the candidates. The annotation and adjustment records, problem analysis records, and guidance records of the previous remote expert are packaged into a standardized information package and transferred to an available target remote expert. It enhances the efficiency of communication, guidance, and the intelligence and efficiency of expert resource utilization in remote MRI collaboration. Attached Figure Description
[0014] Figure 1 This is a block diagram of the remote collaboration and expert guidance system for nuclear magnetic resonance provided in the embodiments of this application.
[0015] Explanation of reference numerals in the attached diagram: 1. Parameter preset module; 2. Image quality module; 3. Data transmission module; 4. Remote adjustment module; 5. Remote guidance module; 6. Expert relay module. Detailed Implementation
[0016] The following is in conjunction with the appendix Figure 1 This application will be described in further detail, but the embodiments of the present invention are not limited thereto.
[0017] This application discloses a remote collaboration and expert guidance system for nuclear magnetic resonance imaging.
[0018] In this embodiment, a remote collaboration and expert guidance system for nuclear magnetic resonance imaging includes: Parameter preset module 1: Used to obtain basic patient information, examination site and previous images, generate scan preset parameters, and remote experts review and adjust the scan preset parameters to obtain the target scan preset parameters; Image quality module 2: It is used to perform MRI examination on patients based on preset parameters of target scan, and to check the image quality during the image generation process, identify image problems, generate self-repair guidelines based on image problems, and perform image processing based on self-repair guidelines to obtain the target MRI image; Data transmission module 3: used to obtain the communication channel between the hospital and remote experts, obtain the channel bandwidth based on the communication channel, perform intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain the core image, and send the core image to the remote experts; Remote adjustment module 4: Used to acquire voice and gesture data from remote experts, recognize the voice and gesture data, obtain real-time control signals and real-time guidance information from remote experts, and adjust the parameters of the MRI equipment according to the real-time control signals; Remote guidance module 5: Based on real-time guidance information, it performs spatial positioning on the real screen of the hospital's operating terminal, generates visual guidance, and guides the operator to perform operations; Expert Relay Module 6: Used to establish a collaborative task queue. The collaborative task queue has multiple remote experts to be selected. When a remote expert leaves, the remote expert's annotation adjustment record, problem analysis record, and guidance record are packaged into a standardized information package and handed over to another idle expert in the collaborative task queue.
[0019] The steps for obtaining basic patient information, examination sites, and previous images, generating preset scanning parameters, and having remote experts review and adjust these preset parameters to arrive at the target preset scanning parameters are as follows: Obtain basic patient information, examination sites, and previous images; generate basic examination parameters based on the patient's basic information. Based on the examination site and previous images, the patient's problem points are located to obtain the lesion location. Based on the lesion location, the scanning depth and scanning area are determined. By combining basic inspection parameters, scanning depth, and scanning area, preset scanning parameters are generated; The preset scanning parameters are sent to a remote expert, who reviews and adjusts them. After obtaining the target preset scanning parameters, the data is transmitted back.
[0020] In this application, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, the system obtains the patient's basic information, including age 55, male, weight 70 kg, and the examination site being the head. It also retrieves the patient's head CT images from three months prior as previous images. Based on this information, the system automatically generates basic examination parameters, such as selecting T1-weighted and T2-weighted scanning sequences and setting the slice thickness to 5 mm. Then, combining the examination site and previous images, the system locates the problem area. By comparing the CT images, a suspected lesion approximately 2 cm in diameter is found in the left temporal lobe region. Based on the location of this lesion, the system determines that the scan depth needs to cover the entire brain, and the scan area is set to 20 cm × 20 cm. Next, the system integrates the basic examination parameters, scan depth, and scan area to generate initial preset scan parameters. These parameters are then sent to Dr. Wang, a remote expert located at a higher-level hospital. After reviewing the data, Dr. Wang determined that a diffusion-weighted imaging sequence needed to be added to better display the lesion features, and the slice thickness was adjusted to 3 millimeters to improve resolution. After the adjustments were completed, the final target scan preset parameters were sent back to the primary hospital.
[0021] The steps for performing image quality checks during image generation to identify image problems, generating a self-repair guide based on these problems, and then processing the image according to the self-repair guide to obtain the target MRI image are as follows: The image generation process is monitored to obtain the image data stream before image generation, and the quality of the image data stream is identified to obtain the data stream quality points and the corresponding problem data streams. Identify the image problem type, location, and phenomenon corresponding to the data stream quality points to obtain the image problem; Based on the image problem, the problem source is traced to obtain the data output process of the problem data flow, and the anomaly monitoring of the data output process is carried out to obtain the target anomaly point; Extract abnormal data from target anomalies and generate a self-repair guide based on the abnormal data output. Based on the self-healing guidelines, an image restoration operation is generated. The image problems are then addressed using the image restoration operation to obtain the target MRI image.
[0022] In practice, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, the system began scanning based on the target scan preset parameters adjusted by remote experts. During image generation, the system monitored the raw K-space data stream in real time, identifying data stream quality through algorithms. It discovered that in a set of T2-weighted image data streams, the signal-to-noise ratio suddenly dropped by 30%, marking this as a problematic data stream. The system further identified the image problem type corresponding to this problematic data stream as motion artifact, located in the central region of the image, manifesting as image blurring and ghosting. Based on this image problem, the system traced the source of the problem, back to the generation process of the problematic data stream, and found that the patient made a slight head movement during the 15-second acquisition, which was marked as the target abnormality. The system extracted the data output operation of this abnormality—i.e., the patient moved during the scan. Based on this operation, the system generated a self-repair guideline: recommending re-acquiring data for this time period and prompting the technician to strengthen patient fixation and communication in subsequent scans. Finally, the system performs image restoration operations according to the self-healing guidelines, removes abnormal data and uses interpolation compensation with neighboring normal data, and finally obtains a clear and usable target MRI image.
[0023] The steps are as follows: First, obtain the channel bandwidth based on the communication channel. Then, perform intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain the core image. Finally, send the core image to a remote expert. When the signal is first established between the hospital and the remote expert, the communication channel between the hospital and the remote expert is obtained, and the bandwidth of the communication channel is tested to obtain the channel bandwidth. Based on the channel bandwidth, the upper limit of the communication rate of the communication channel is obtained, and the redundancy is set according to the upper limit of the communication rate. The channel bandwidth is then set according to the redundancy to obtain the target channel bandwidth. Obtain the image data volume of the target NMR image, and based on the image data volume, determine the minimum channel bandwidth required for the target NMR image; Determine whether the minimum channel bandwidth is greater than the target channel bandwidth; If it is determined that the minimum channel bandwidth is greater than the target channel bandwidth, the target NMR image is identified to obtain key image information; Based on key image information, the target MRI image is cropped to obtain the core image, which is then sent to a remote expert.
[0024] In practice, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, after establishing a connection with remote expert Dr. Wang, the system automatically tested the communication channel between the two hospitals, measuring the current channel bandwidth to be 50Mbps. Based on this bandwidth, the system calculated the upper limit of the communication rate for this channel and reserved a 10% redundancy, setting the target channel bandwidth to 45Mbps. Next, the system acquired the newly generated target MRI image, which had a data size of 800MB. Calculations showed that the minimum channel bandwidth required for smooth transmission of the entire image was approximately 100Mbps. The system determined that the minimum channel bandwidth was greater than the target channel bandwidth. Therefore, the system intelligently identified the target MRI image, using an image segmentation algorithm combined with lesion localization information to identify the lesion in the left temporal lobe region and its surrounding 3cm area as key image information. Then, based on this key image information, the system intelligently cropped the original whole-brain image, retaining only the core region image centered on the lesion and measuring 10cm x 10cm, resulting in a core image with a data size of only 120MB, which was then prioritized and sent to remote expert Dr. Wang for review.
[0025] The steps for sending the core image to a remote expert are as follows: During the transmission of core images, the network quality of the communication channel is monitored to obtain real-time network quality parameters; Determine whether the real-time network quality parameters are consistently lower than the preset standard network quality parameters within a preset time period; If it is determined that the network quality parameters are consistently lower than the preset standard network quality parameters within a preset time period, then extract the key data information from the core image. The key data information is packaged and compressed to generate a key data packet, which is then sent to a remote expert. Extract secondary data information from the core image that is not required in real time, in addition to the key data information, and resume the transmission of the secondary data information from the breakpoint.
[0026] In practice, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, the system continuously monitored the network quality of the communication channel during the transmission of the core image to Dr. Wang. Five seconds after transmission began, the system detected fluctuations in real-time network quality parameters (such as latency and packet loss rate), which remained below the preset stable transmission standard for the next 10 seconds. The system determined that the network quality remained poor. Therefore, the system immediately extracted the most critical data from the transmitted core image data, such as the edge contour data of the lesion area, grayscale histogram statistical features, and preliminary texture analysis results. These data were very small. The system packaged and compressed this critical data into a single 5MB key data packet, which was then prioritized and sent to Dr. Wang so that he could make a basic assessment of the lesion. Simultaneously, the system extracted the remaining non-real-time secondary data from the core image, such as smooth areas of the image background, marked this data, and initiated a breakpoint resume mechanism, waiting for network quality to recover before resuming transmission.
[0027] The steps for acquiring voice and gesture data from remote experts, recognizing the voice and gesture data to obtain real-time control signals and guidance information from the remote experts, and adjusting the parameters of the MRI equipment based on the real-time control signals are as follows: Acquire voice and gesture data from remote experts; Based on the speech data, speech information recognition is performed on the speech data to obtain text content, and keyword extraction and context recognition are performed on the text content to obtain language instruction information; Based on gesture data, gesture information recognition is performed on the gesture data to obtain hand posture information, posture change information, and hand trajectory information; Based on hand posture information, posture change information, and hand trajectory information, the gesture command information of the remote expert is obtained; By combining language command information and gesture command information, the command information is collected and divided to obtain real-time control signals and real-time guidance information. The real-time control signals are transmitted back to the hospital's MRI equipment to adjust the parameters of the MRI equipment.
[0028] In practice, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, remote expert Dr. Wang needs to provide remote guidance while reviewing the images. The system acquires Dr. Wang's voice and gesture data through his microphone and camera. The voice data is Dr. Wang saying, "Increase the TR time of the T1 sequence to 800 milliseconds." The system recognizes this voice, converts it to text, and extracts the keywords "T1 sequence," "TR," "increase," and "800 milliseconds." Combined with the context, it generates the language instruction: Adjust the repetition time parameter of the T1-weighted sequence to 800ms. Simultaneously, the camera captures Dr. Wang making a rightward swipe gesture. The system recognizes this gesture data, finding the hand posture to be an outstretched palm, and the posture change to be a smooth movement from the left to the right of the screen along a horizontal straight line. Based on this information, the system generates the gesture instruction: Turn to the next sequence image. The system then combines and categorizes the verbal and gesture commands, classifying "adjusting the TR time of the T1-weighted sequence to 800ms" as a real-time control signal requiring immediate execution, while "viewing the next image" is categorized as real-time guidance for the operator. Finally, the system directly transmits the "adjusting the TR time to 800ms" real-time control signal back to the MRI equipment in the primary hospital, where the equipment automatically adjusts the parameter.
[0029] Based on real-time guidance information, spatial positioning is performed on the actual screen of the hospital's operating terminal to generate visual guidance and guide the operator through the operation steps, specifically: The system acquires real-time images from the hospital's control panel, identifies these images, and obtains multiple control points and control areas. Based on real-time guidance information, multiple operation points and operation areas are matched to obtain the target operation point and target operation area; Extract the actual spatial location of the target operation point and the target operation area, as well as the corresponding operation action; Based on the actual spatial location and operational actions, visual guidance is generated and projected onto the AR display device preset on the operating end.
[0030] In practice, taking an MRI scan of a suspected brain tumor patient at a primary care hospital as an example, the system received real-time guidance from Dr. Wang: "Please adjust the position of the patient's head coil to better fit above the left ear." First, the system acquires real-time monitoring footage from the hospital's control panel. Using image recognition technology, it identifies multiple operation points and areas, including the MRI equipment's control panel, the patient's head, the head coil, and the laser positioning light. Then, the system matches "head coil" and "above the left ear" from the real-time guidance with the identified operation areas, determining that the target operation point is the head coil above the patient's left ear, and the target operation area is the head coil's fixing clip. Next, the system uses augmented reality spatial positioning technology to extract the precise three-dimensional coordinates of the target operation point (the head coil above the left ear) in actual space, along with the corresponding operation action: "rotate the clip for fine-tuning." Finally, based on this actual spatial location and operation, the system generates a visual guide: a flashing green arrow virtual icon, precisely superimposed on the position of the head coil buckle above the patient's left ear displayed on the operator's AR glasses, accompanied by the text prompt "rotate counterclockwise 15 degrees", directly guiding the operator to complete the adjustment operation.
[0031] The steps for establishing a collaborative task queue with multiple remote experts in reserve, and for transferring a standardized information package to another available expert in the collaborative task queue when a remote expert leaves the queue, are as follows: Establish a collaborative task queue, which contains multiple remote candidate experts, and indicate the current idle status of each remote candidate expert. When a remote expert leaves, the system automatically retrieves the expert's annotation and adjustment records, problem analysis records, and guidance records, and generates a standardized information package. Based on the standardized information package, identify the current actual inspection progress and generate a progress docking interface accordingly. Find a remote candidate expert whose current status is marked as idle in the collaborative task queue in order, and mark it as the target remote expert; The standardized information package and progress docking interface are sent to the target remote experts, and the standardized information package is extracted and simplified to generate a quick reading report.
[0032] In practice, taking an MRI scan of a suspected brain tumor patient at a primary hospital as an example, before the remote collaboration began, the system had established a collaborative task queue listing three potential remote experts: Dr. Wang (currently busy), Dr. Li (available), and Dr. Zhang (available), with their statuses marked. Halfway through the examination, Dr. Wang temporarily left due to an urgent meeting. The system automatically detected Dr. Wang's disconnection and immediately captured and packaged all his records from the collaboration—including his previous annotations and adjustments to scan parameters (such as modifying slice thickness), analysis of image problems (such as pointing out artifact types), and operational guidance to the technician (such as adjusting coils)—generating a structured, standardized information package. Based on this information package, the system identified the current examination progress as "T1 and T2 sequence scans completed, preparing for diffusion-weighted sequence." Then, the system generated a progress docking interface, linking to the upcoming scan step. Next, the system searched the collaborative task queue sequentially, finding the first expert marked as "available," Dr. Li, and marked him as the target remote expert. Finally, the system sends the standardized information package and progress docking interface to Dr. Li, extracts and simplifies the key information in the package, and generates a "quick reading report" so that Dr. Li can understand the background in a short time and seamlessly take over the subsequent guidance work.
[0033] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.
Claims
1. A remote collaboration and expert guidance system for nuclear magnetic resonance imaging, characterized in that, include: Parameter preset module: used to acquire basic patient information, examination site and previous images, generate scan preset parameters, and remote experts review and adjust the scan preset parameters to obtain the target scan preset parameters; Image quality module: used to perform MRI examination on the patient based on the preset parameters of the target scan, and to perform image quality check during the image generation process to identify image problems, generate self-repair guidelines based on the image problems, and perform image processing according to the self-repair guidelines to obtain the target MRI image; Data transmission module: used to acquire the communication channel between the hospital and remote experts, obtain the channel bandwidth based on the communication channel, perform intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain the core image, and send the core image to the remote experts; Remote adjustment module: used to acquire voice data and gesture data of remote experts, recognize the voice data and gesture data to obtain real-time control signals and real-time guidance information from remote experts, and adjust the parameters of the MRI equipment according to the real-time control signals; Remote guidance module: Based on the real-time guidance information, it performs spatial positioning on the real screen of the hospital's operating terminal, generates visual guidance, and guides the operator to perform operations; Expert relay module: Used to establish a collaborative task queue, which contains multiple remote experts. When a remote expert leaves, the module packages the remote expert's annotation adjustment records, problem analysis records, and guidance records into a standardized information package and transfers it to another available expert in the collaborative task queue.
2. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 1, characterized in that, The steps of obtaining basic patient information, examination sites, and previous images, generating preset scanning parameters, and having remote experts review and adjust these preset scanning parameters to obtain the target preset scanning parameters are as follows: Obtain basic patient information, examination sites, and previous images; generate basic examination parameters based on the patient's basic information. Based on the examined area and the previous images, the patient's problem points are located to obtain the lesion location. Based on the lesion location, the scanning depth and scanning area are determined. By combining the basic inspection parameters, the scanning depth, and the scanning area, preset scanning parameters are generated; The preset scanning parameters are sent to a remote expert, who reviews and adjusts them to obtain the target preset scanning parameters before sending the data back.
3. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 2, characterized in that, The steps for performing image quality checks during image generation to identify image problems, generating self-repair guidelines based on these problems, and then processing the image according to these guidelines to obtain the target MRI image are as follows: The image generation process is monitored to obtain the image data stream before image generation, and the image data stream is quality identified to obtain data stream quality points and the problem data streams corresponding to the data stream quality points; Identify the image problem type, image problem location, and image problem phenomenon corresponding to the data stream quality point to obtain the image problem; Based on the image problem, the problem source is traced to obtain the data output process of the problem data stream, and the anomaly monitoring of the data output process is performed to obtain the target anomaly point; Extract abnormal data from the target anomaly point and generate a self-repair guide based on the abnormal data output operation; Based on the self-healing guidelines, an image repair operation is generated, and the image problem is processed according to the image repair operation to obtain the target MRI image.
4. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 3, characterized in that, The steps of obtaining the channel bandwidth based on the communication channel, performing intelligent image segmentation on the target MRI image based on the channel bandwidth to obtain the core image, and sending the core image to a remote expert are as follows: When the signal is first established between the hospital and the remote expert, the communication channel between the hospital and the remote expert is obtained, and the bandwidth of the communication channel is tested to obtain the channel bandwidth. Based on the channel bandwidth, the upper limit of the communication rate of the communication channel is obtained, and a redundancy amount is set according to the upper limit of the communication rate. The channel bandwidth is then set according to the redundancy amount to obtain the target channel bandwidth. The image data volume of the target NMR image is obtained, and the minimum channel bandwidth required for the target NMR image is obtained based on the image data volume; Determine whether the minimum channel bandwidth is greater than the target channel bandwidth; If it is determined that the minimum channel bandwidth is greater than the target channel bandwidth, then the target MRI image is identified to obtain key image information; Based on the key image information, the target MRI image is cropped to obtain a core image, which is then sent to a remote expert.
5. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 4, characterized in that, The steps for sending the core image to a remote expert are as follows: During the transmission of the core image, the network quality of the communication channel is monitored to obtain real-time network quality parameters; Determine whether the real-time network quality parameter is consistently lower than the preset standard network quality parameter within a preset time period; If it is determined that the network quality parameter is consistently lower than the preset standard network quality parameter within a preset time period, then key data information in the core image is extracted. The key data information is packaged and compressed to generate a key data packet, which is then sent to a remote expert. Extract secondary data information that is not required in real time from the core image, in addition to the key data information, and resume transmission of the secondary data information from breakpoints.
6. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 5, characterized in that, The steps of acquiring voice and gesture data from a remote expert, recognizing the voice and gesture data to obtain real-time control signals and guidance information from the remote expert, and adjusting the parameters of the MRI equipment based on the real-time control signals are as follows: Acquire voice and gesture data from remote experts; Based on the voice data, voice information recognition is performed on the voice data to obtain text content, and keyword extraction and context recognition are performed on the text content to obtain language instruction information; Based on the gesture data, gesture information recognition is performed on the gesture data to obtain hand posture information, posture change information, and hand trajectory information; Based on the hand posture information, the posture change information, and the hand trajectory information, the gesture command information of the remote expert is obtained; By combining the language instruction information and the gesture instruction information, the instruction information is collected and divided to obtain real-time control signals and real-time guidance information. The real-time control signal is transmitted back to the hospital's MRI equipment to adjust the parameters of the MRI equipment.
7. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 6, characterized in that, Based on the real-time guidance information, spatial positioning is performed on the actual screen of the hospital's operating terminal to generate visual guidance, which guides the operator through the following steps: The system acquires real-time images from the hospital's control panel, identifies these images, and obtains multiple control points and control areas. Based on the real-time guidance information, multiple operation points and operation areas are matched to obtain target operation points and target operation areas; Extract the actual spatial location of the target operation point and the target operation area, as well as the corresponding operation action; Based on the actual spatial location and the operation action, a visual guide is generated and projected onto an AR display device preset on the operation terminal.
8. The remote collaboration and expert guidance system for nuclear magnetic resonance imaging according to claim 7, characterized in that, The steps for establishing a collaborative task queue, in which multiple remote experts are selected, and when a remote expert leaves, to package their annotation adjustment records, problem analysis records, and guidance records into a standardized information package and transfer it to another available expert in the collaborative task queue, are as follows: Establish a collaborative task queue, in which multiple remote candidate experts are set, and the current idle status of each remote candidate expert is marked; When a remote expert leaves, the system automatically retrieves the expert's annotation and adjustment records, problem analysis records, and guidance records, and generates a standardized information package. Based on the standardized information package, the current actual inspection progress is identified, and a progress docking interface is generated based on the actual inspection progress. In the collaborative task queue, a remote candidate expert whose current status is marked as idle is found in sequence and marked as the target remote expert; The standardized information package and the progress docking interface are sent to the target remote expert, and the standardized information package is extracted and simplified to generate a quick reading report.