A portable magnetic resonance system

Abstract

The application provides a portable magnetic resonance system, and relates to the technical field of magnetic resonance imaging, which comprises: a data acquisition device, which is used for acquiring real-time electromagnetic interference data under the current environment of the portable magnetic resonance device before the portable magnetic resonance device starts scanning; an image processing module, which is used for acquiring a real-time magnetic resonance image scanned under the current environment of the magnetic resonance imaging, and inputting the real-time magnetic resonance image and the real-time electromagnetic interference data into a pre-trained image optimization model to obtain an optimized magnetic resonance image; and an image transmission module, which is used for sending the real-time magnetic resonance image and / or the optimized magnetic resonance image to a doctor terminal for comparison and viewing by the doctor. The beneficial effect is that the magnetic resonance image scanned by the portable magnetic resonance device can be collected and optimized based on the real-time electromagnetic interference data, so that the magnetic resonance device can provide high-quality magnetic resonance images while being portable, and the medical needs can be met.

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance imaging technology, and more particularly to a portable magnetic resonance system. Background Technology

[0002] With the continuous advancement of medical imaging technology, patients expect a better experience during diagnosis, hospitals and research institutions desire more cost-effective diagnostic equipment, and radiology departments and researchers want guaranteed image quality. However, existing MRI equipment typically requires placement in a shielded room. Whether hospitalized or seeking medical attention, patients need to descend to the first floor or a basement level of the shielded room, remove various metal accessories, and then be pushed into a narrow tubular space while lying down for the MRI examination. This is very inconvenient for critically ill patients or those with limited mobility. As a result, portable MRI scanners have emerged.

[0003] Portable magnetic resonance imaging (MRI) equipment represents another innovative development in magnetic resonance technology. As a supplement and extension to traditional large-scale conventional NMR equipment, it features low field strength, small size, portability, safety without radiation, and ease of operation. This testing device can be moved to intensive care units (ICUs), operating rooms, emergency rooms, or regular wards, allowing for bedside examinations. It provides efficient and accurate diagnostic data for clinicians, helping them to understand changes in patients' conditions in a timely manner, adopt appropriate treatment plans, reduce the inconvenience caused by transporting patients, effectively avoid unexpected medical risks and adverse events, and expand the scope of clinical applications of NMR.

[0004] Although portable magnetic resonance imaging (MRI) devices have relatively simple requirements for the external environment and do not require special shielding, the image quality is still inevitably affected by many uncontrollable environmental factors due to their operation in the external environment. Summary of the Invention

[0005] To address the problems existing in the prior art, the present invention provides a portable magnetic resonance system, comprising:

[0006] An electromagnetic interference detection device is installed on a portable magnetic resonance imaging (MRI) device to acquire real-time electromagnetic interference data of the environment in which the portable MRI device is currently located before the portable MRI device begins scanning.

[0007] An image processing module, connected to the electromagnetic interference detection device, is used to acquire the real-time magnetic resonance image obtained by the magnetic resonance imaging in the current environment, and input the real-time magnetic resonance image and the real-time electromagnetic interference data into a pre-trained image optimization model to obtain an optimized magnetic resonance image.

[0008] An image transmission module, connected to the image processing module and the doctor's terminal respectively, is used to send the real-time magnetic resonance image and / or the optimized magnetic resonance image to the doctor's terminal for comparison and viewing by the doctor.

[0009] Preferably, the image processing module includes:

[0010] The model storage unit is used to store the image optimization model associated with different preset electromagnetic interference data;

[0011] A model selection unit, connected to the model storage unit, is used to select the image optimization model associated with the preset electromagnetic interference data that matches the real-time electromagnetic interference data as the current optimization model;

[0012] An image processing unit, connected to the model selection unit, is used to input the real-time magnetic resonance image and the real-time electromagnetic interference data into the current optimization model to obtain the optimized magnetic resonance image.

[0013] Preferably, the preset electromagnetic interference data includes multiple electromagnetic interference intensity ranges and their associated electromagnetic interference phase ranges; then the model selection unit includes:

[0014] A matching subunit is used to obtain the closest electromagnetic interference intensity range and associated electromagnetic interference phase based on the real-time interference intensity and real-time interference phase contained in the real-time electromagnetic interference data.

[0015] Select a subunit and connect it to the matching subunit, which is used to select the image optimization model that is closest to the electromagnetic interference intensity range and associated with the electromagnetic interference phase as the current optimization model.

[0016] Preferably, the electromagnetic interference detection device includes:

[0017] Multiple interference collectors are arrayed around the portable magnetic resonance imaging device to collect electromagnetic interference signals in the current environment of the portable magnetic resonance imaging device.

[0018] A signal processor, connected to each of the interference acquisition units, is used to perform signal strength processing and signal phase processing on the acquired electromagnetic interference signals to obtain real-time interference intensity and real-time interference phase, which are then used as the real-time electromagnetic interference data.

[0019] Preferably, the signal processor and each of the acquisition sensors are connected via shielded cables.

[0020] Preferably, the image optimization model includes a feature prediction sub-model and an image generation adversarial sub-model connecting the feature prediction sub-model; then the image processing unit includes:

[0021] The first processing subunit is used to input the real-time electromagnetic interference data into the feature prediction sub-model to predict the magnetic resonance imaging difference features generated by the real-time electromagnetic interference data.

[0022] The second processing subunit, connected to the first processing subunit, is used to input the magnetic resonance imaging difference features and the real-time magnetic resonance image into the image generation adversarial model to generate the optimized magnetic resonance image after removing the magnetic resonance imaging difference features.

[0023] Preferably, it further includes a model training module connected to the image processing module, the model training module comprising:

[0024] The first acquisition unit is used to acquire several first magnetic resonance images obtained by magnetic resonance scanning when the portable magnetic resonance device is placed in a shielded room.

[0025] The second acquisition unit is used to acquire several second magnetic resonance images obtained by the portable magnetic resonance device under different medical environments for magnetic resonance scanning.

[0026] The feature comparison unit is connected to the first acquisition unit and the second acquisition unit respectively. It is used to perform feature comparison on each of the first magnetic resonance images with the same scanning parameters and the corresponding second magnetic resonance images to obtain corresponding image difference features, and associate the image difference features with the corresponding preset electromagnetic interference data in the medical environment.

[0027] The model training unit, connected to the feature comparison unit, is used to perform transfer training on the pre-trained large language model to obtain a feature prediction sub-model with the preset electromagnetic interference data as input and the magnetic resonance imaging difference features as output, and an image generation adversarial sub-model with the magnetic resonance imaging difference features and the second magnetic resonance image as input and the first magnetic resonance image as output, and saves it to the image processing module.

[0028] The above technical solution has the following advantages or beneficial effects: it can collect real-time electromagnetic interference data in the medical environment where the portable magnetic resonance imaging device is located during scanning, and optimize the magnetic resonance images obtained by the portable magnetic resonance imaging device based on the real-time electromagnetic interference data, so that the magnetic resonance imaging device can provide high-quality magnetic resonance images while being portable, thus meeting medical needs. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the structure of a portable magnetic resonance system, which is a preferred embodiment of the present invention. Detailed Implementation

[0030] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The present invention is not limited to this embodiment; other embodiments that conform to the spirit of the present invention may also fall within the scope of the present invention.

[0031] In a preferred embodiment of the present invention, based on the above-mentioned problems existing in the prior art, a portable magnetic resonance system is provided, such as... Figure 1 As shown, it includes:

[0032] Electromagnetic interference detection device 1 is installed on a portable magnetic resonance imaging device and is used to acquire real-time electromagnetic interference data of the environment in which the portable magnetic resonance imaging device is currently located before the portable magnetic resonance imaging device starts scanning.

[0033] Image processing module 2 is connected to electromagnetic interference detection device 1. It is used to acquire real-time magnetic resonance images obtained by scanning in the current environment, and input the real-time magnetic resonance images and real-time electromagnetic interference data into a pre-trained image optimization model to obtain optimized magnetic resonance images.

[0034] Image transmission module 3 is connected to image processing module 2 and doctor terminal 4 respectively, and is used to send real-time magnetic resonance images and / or optimized magnetic resonance images to doctor terminal 4 for doctors to compare and view.

[0035] Specifically, portable MRI machines, due to their small size and portability, can be easily moved into wards, emergency rooms, and ICUs to facilitate timely MRI scans for patients who are unable to move. However, wards and emergency rooms typically experience high traffic and are equipped with numerous monitoring instruments. Electromagnetic interference generated by both personnel and monitoring instruments is random and uncontrollable, and different external environments can affect the quality of MRI images to varying degrees. Therefore, in this embodiment, an electromagnetic interference detection device 1 is installed on the portable MRI machine to acquire real-time electromagnetic interference data in the medical environment in which the portable MRI machine operates. This real-time electromagnetic interference data is then used as one of the inputs to an image optimization model to optimize the real-time MRI image, minimizing the impact of real-time electromagnetic interference data on the MRI imaging.

[0036] After processing and optimizing the MRI images, they can be displayed in real-time on the screen of a portable MRI device, allowing the scanning physician to view them promptly. If transmission to other physicians is required, the images can be remotely sent to the corresponding physician's terminal 4 via the image transmission module 3, enabling remote diagnosis or remote assistance. Furthermore, when network conditions permit, both real-time and optimized MRI images can be sent to the physician's terminal 4, allowing for comparison and a more comprehensive assessment of the patient's condition. In cases of poor network conditions, only the optimized or real-time MRI images can be sent to the corresponding physician's terminal 4, depending on the physician's needs. The image transmission module 3 includes, but is not limited to, transmission methods such as Bluetooth, Wi-Fi, or 5G.

[0037] Furthermore, based on the above analysis, the impact weight of electromagnetic interference on magnetic resonance imaging varies under different medical environments. For example, when the electromagnetic interference intensity is low, it has little impact on the image quality of portable magnetic resonance imaging; the image may exhibit slight noise or artifacts, but it will not significantly affect the diagnosis. When the electromagnetic interference intensity increases, it may cause problems such as image artifacts, noise, and geometric distortion, reducing image contrast, blurring details, or even making it impossible to identify the target structure at all. This will seriously affect the doctor's interpretation and diagnosis of the image. Similarly, the phase of electromagnetic interference can cause phase shift during the imaging process, leading to image blurring, distortion, or artifacts. To improve the image optimization effect of the image optimization model, it is preferable to train the corresponding image optimization model with different preset electromagnetic interference data. In a preferred embodiment of the present invention, the image processing module 2 includes:

[0038] Model storage unit 21 is used to store the image optimization model associated with different preset electromagnetic interference data;

[0039] The model selection unit 22 is connected to the model storage unit 21 and is used to select the image optimization model associated with the preset electromagnetic interference data that matches the real-time electromagnetic interference data as the current optimization model.

[0040] The image processing unit 23 is connected to the model selection unit 22 and is used to input real-time magnetic resonance images and real-time electromagnetic interference data into the current optimization model to obtain optimized magnetic resonance images.

[0041] Specifically, in this embodiment, for matching the image optimization model, the portable magnetic resonance imaging device is equipped with an electromagnetic interference detection device 1, which includes:

[0042] Multiple interference collectors 11 are arrayed around the portable magnetic resonance device to collect electromagnetic interference signals in the current environment of the portable magnetic resonance device.

[0043] The signal processor 12 is connected to each interference collector 11 and is used to process the collected electromagnetic interference signals by signal strength processing and signal phase processing to obtain real-time interference strength and real-time interference phase, which are then used as real-time electromagnetic interference data.

[0044] Specifically, in this embodiment, the interference collector 11 is preferably a wideband receiver to receive interference signals in various frequency ranges. It may also include electromagnetic radiation sensors, electromagnetic induction sensors, electromagnetic interference monitoring sensors, etc. There are no limitations here, as long as it can collect electromagnetic interference signals.

[0045] The acquired electromagnetic interference signals are processed by the signal processor 12. Signal filtering can be performed before signal strength and phase processing. Specifically, the signal filtering can be designed according to the design parameters of the portable magnetic resonance imaging (MRI) device. For example, electromagnetic interference signals that the portable MRI device can shield itself can be filtered out using appropriate filters, while electromagnetic interference signals that the portable MRI device cannot shield itself are retained, thus reducing the amount of data processing. More preferably, the signal processor 12 can be implemented using an FPGA chip.

[0046] In a preferred embodiment of the present invention, the signal processor 12 and each acquisition sensor 11 are connected by shielded cables to prevent additional electromagnetic interference.

[0047] In a preferred embodiment of the present invention, the image optimization model includes a feature prediction sub-model and an image generative adversarial sub-model connecting the feature prediction sub-model; then the image processing unit 23 includes:

[0048] The first processing subunit 231 is used to input real-time electromagnetic interference data into the feature prediction submodel to predict the magnetic resonance imaging difference features generated by the real-time electromagnetic interference data.

[0049] The second processing subunit 232, connected to the first processing subunit 231, is used to generate an adversarial model from the input image of magnetic resonance imaging differential features and real-time magnetic resonance image to generate an optimized magnetic resonance image after removing magnetic resonance imaging differential features.

[0050] In a preferred embodiment of the present invention, a model training module 5 is further included, connected to the image processing module 2. The model training module 5 includes:

[0051] The first acquisition unit 51 is used to acquire several first magnetic resonance images obtained by magnetic resonance scanning when the portable magnetic resonance equipment is placed in a shielded room.

[0052] The second acquisition unit 52 is used to acquire several second magnetic resonance images obtained by placing a portable magnetic resonance device in different medical environments for magnetic resonance scanning.

[0053] The feature comparison unit 53 is connected to the first acquisition unit 51 and the second acquisition unit 52 respectively. It is used to perform feature comparison on each first magnetic resonance image with the same scanning parameters and the corresponding second magnetic resonance image to obtain the corresponding image difference features, and associate the image difference features with the corresponding preset electromagnetic interference data in the medical environment.

[0054] The model training unit 54 is connected to the feature comparison unit 53, which is used to perform transfer training on the pre-trained large language model to obtain a feature prediction sub-model with preset electromagnetic interference data as input and magnetic resonance imaging difference features as output, and an image generation adversarial sub-model with magnetic resonance imaging difference features and a second magnetic resonance image as input and a first magnetic resonance image as output, and saves it to the image processing module 2.

[0055] Specifically, in this embodiment, the second magnetic resonance image can be obtained by scanning during actual use of the portable magnetic resonance device, or by simulating different medical environments to collect the corresponding second magnetic resonance image. After collecting enough first magnetic resonance images and corresponding second magnetic resonance images, model training can be performed.

[0056] The preferred large-scale prediction model is the ChatGPT model, with the GPT-4 large-scale language model being even more preferred. The GPT-4 model can generate response text for multimodal inputs of text and images, as well as classify, analyze, and extract implicit semantics from visual elements, demonstrating excellent response capabilities. Preferably, a feature prediction sub-model is obtained through transfer learning training utilizing the semantic analysis capabilities of the GPT-4 model, and an image generation adversarial sub-model is obtained through transfer learning training utilizing the generative capabilities of the GPT-4 model.

[0057] Furthermore, the aforementioned model training module 5 is preferably configured in a remote server to reduce the computing power requirements of the portable magnetic resonance device, thereby reducing the device cost.

[0058] The above description is merely a preferred embodiment of the present invention and does not limit the implementation and protection scope of the present invention. Those skilled in the art should realize that any equivalent substitutions and obvious changes made using the content of this specification and illustrations should be included within the protection scope of the present invention.

Claims

1. A portable magnetic resonance system, characterized in that, include: An electromagnetic interference detection device is installed on a portable magnetic resonance imaging (MRI) device to acquire real-time electromagnetic interference data of the environment in which the portable MRI device is currently located before the portable MRI device begins scanning. An image processing module, connected to the electromagnetic interference detection device, is used to acquire real-time magnetic resonance images scanned by the portable magnetic resonance device in the current environment, and input the real-time magnetic resonance images and the real-time electromagnetic interference data into a pre-trained image optimization model to obtain an optimized magnetic resonance image. An image transmission module is connected to the image processing module and the doctor's terminal, respectively, and is used to send the real-time magnetic resonance image and / or the optimized magnetic resonance image to the doctor's terminal for comparison and viewing by the doctor; The image processing module includes: The model storage unit is used to store the image optimization model associated with different preset electromagnetic interference data; A model selection unit, connected to the model storage unit, is used to select the image optimization model associated with the preset electromagnetic interference data that matches the real-time electromagnetic interference data as the current optimization model; An image processing unit, connected to the model selection unit, is used to input the real-time magnetic resonance image and the real-time electromagnetic interference data into the current optimization model to obtain the optimized magnetic resonance image.

2. The portable magnetic resonance system according to claim 1, characterized in that, The preset electromagnetic interference data includes multiple electromagnetic interference intensity ranges and their associated electromagnetic interference phase ranges; therefore, the model selection unit includes: A matching subunit is used to obtain the closest electromagnetic interference intensity range and associated electromagnetic interference phase based on the real-time interference intensity and real-time interference phase contained in the real-time electromagnetic interference data. Select a subunit and connect it to the matching subunit, which is used to select the image optimization model that is closest to the electromagnetic interference intensity range and associated with the electromagnetic interference phase as the current optimization model.

3. The portable magnetic resonance system according to claim 1, characterized in that, The electromagnetic interference detection device includes: Multiple interference collectors are arrayed around the portable magnetic resonance imaging device to collect electromagnetic interference signals in the current environment of the portable magnetic resonance imaging device. A signal processor, connected to each of the interference acquisition units, is used to perform signal strength processing and signal phase processing on the acquired electromagnetic interference signals to obtain real-time interference intensity and real-time interference phase, which are then used as the real-time electromagnetic interference data.

4. The portable magnetic resonance system according to claim 3, characterized in that, The signal processor and each of the interference collectors are connected by shielded cables.

5. The portable magnetic resonance system according to claim 1, characterized in that, The image optimization model includes a feature prediction sub-model and an image generation adversarial sub-model connecting the feature prediction sub-model; therefore, the image processing unit includes: The first processing subunit is used to input the real-time electromagnetic interference data into the feature prediction sub-model to predict the magnetic resonance imaging difference features generated by the real-time electromagnetic interference data. The second processing subunit, connected to the first processing subunit, is used to input the magnetic resonance imaging difference features and the real-time magnetic resonance image into the image generation adversarial model to generate the optimized magnetic resonance image after removing the magnetic resonance imaging difference features.

6. The portable magnetic resonance system according to claim 5, characterized in that, It also includes a model training module connected to the image processing module, the model training module comprising: The first acquisition unit is used to acquire several first magnetic resonance images obtained by magnetic resonance scanning when the portable magnetic resonance device is placed in a shielded room. The second acquisition unit is used to acquire several second magnetic resonance images obtained by the portable magnetic resonance device under different medical environments for magnetic resonance scanning. The feature comparison unit is connected to the first acquisition unit and the second acquisition unit respectively. It is used to perform feature comparison on each of the first magnetic resonance images with the same scanning parameters and the corresponding second magnetic resonance images to obtain corresponding image difference features, and associate the image difference features with the corresponding preset electromagnetic interference data in the medical environment. The model training unit, connected to the feature comparison unit, is used to perform transfer training on the pre-trained large language model to obtain a feature prediction sub-model with the preset electromagnetic interference data as input and the magnetic resonance imaging difference features as output, and an image generation adversarial sub-model with the magnetic resonance imaging difference features and the second magnetic resonance image as input and the first magnetic resonance image as output, and saves it to the image processing module.

Images