Nuclear magnetic inspection digital auxiliary device

Abstract

The invention provides a nuclear magnetic examination digital auxiliary device, which comprises a data acquisition module, a preprocessing module, an AI large model analysis module and a data enhancement module, and is used for acquiring NMR spectrogram data; preprocessing the NMR spectrogram data to obtain preprocessed NMR spectrogram data; performing automatic analysis on the preprocessed NMR spectrogram data by using a pre-trained AI large model to obtain analyzed NMR spectrogram data; and generating diversified NMR sample data based on the analyzed NMR spectrogram data, and matching the NMR sample data with a nuclear magnetic examination report to complete standardization of the NMR spectrogram data and the nuclear magnetic examination report. According to the technical scheme, the analysis efficiency and accuracy of nuclear magnetic inspection are improved, data enhancement is carried out at the same time, and the accuracy, standardization and reliability of nuclear magnetic inspection are improved; and the demand and dependence on professional doctors are reduced.

Description

Background Art

[0002] With the improvement of people's health awareness and the aging of the population, the demand for MRI examinations has gradually increased, while the number of MRI doctors is extremely short. This has resulted in MRI doctors often being overloaded with work, but patients still need to wait for a while before getting the report results. If the MRI doctor's level is not good, misdiagnosis or missed diagnosis may occur. In addition, excessive workload may also cause fatigue of MRI doctors, resulting in misdiagnosis or missed diagnosis, which will bring a series of adverse consequences to the patient's clinical treatment. Summary of the invention

[0003] The present application provides a digital auxiliary device for nuclear magnetic resonance examination, which is used to automatically analyze NMR data through an AI large model to improve the analysis efficiency and accuracy, and at the same time perform data enhancement to improve the accuracy, standardization and reliability of nuclear magnetic resonance examination.

[0004] In a first aspect, the present application provides a nuclear magnetic resonance standardization method, comprising the following steps:

[0005] Obtain NMR spectrum data;

[0006] Preprocessing the NMR spectrum data to obtain preprocessed NMR spectrum data;

[0007] Automatically parsing the preprocessed NMR spectrum data using a pretrained AI large model to obtain parsed NMR spectrum data;

[0008] Based on the analyzed NMR spectrum data, diversified NMR sample data are generated, and the NMR sample data are matched with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0009] In the above technical scheme, NMR spectrum data is acquired; the NMR spectrum data is preprocessed to obtain preprocessed NMR spectrum data; the preprocessed NMR spectrum data is automatically parsed using a pretrained AI large model to obtain parsed NMR spectrum data; diversified NMR sample data is generated based on the parsed NMR spectrum data, and the NMR sample data is matched with the nuclear magnetic resonance examination report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report; the analysis efficiency and accuracy are improved, and data enhancement is performed at the same time to improve the accuracy, standardization and reliability of nuclear magnetic resonance examination; the demand and dependence on professional doctors are reduced.

[0010] In a specific implementation scheme, the step of preprocessing the NMR spectrum data to obtain the preprocessed NMR spectrum data is specifically:

[0011] The NMR spectrum data is cleaned and denoised to obtain pre-processed NMR spectrum data.

[0012] In a specific implementation scheme, the pre-processed NMR spectrum data is automatically parsed using a pre-trained AI large model to obtain the steps of parsing the NMR spectrum data, specifically:

[0013] First, the NMR spectra and their corresponding molecular structure data are used for training to learn the mapping relationship between the NMR spectra and the molecular structure;

[0014] During the analysis process, the pre-processed NMR spectrum data is input, and the corresponding analyzed NMR spectrum data is automatically output using the AI ​​large model.

[0015] In a specific embodiment, the analyzed NMR spectrum data includes molecular structures and their probability rankings.

[0016] In a specific implementation scheme, the steps of generating diversified NMR sample data based on the analyzed NMR spectrum data, matching the NMR sample data with the nuclear magnetic resonance examination report, and completing the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report are specifically as follows:

[0017] Based on the analyzed NMR spectrum data, each peak in the spectrum is identified and its position and intensity are determined;

[0018] Use fitting methods to process overlapping peaks or peaks with complex shapes and obtain peak parameters as data analysis results;

[0019] Generating diversified NMR sample data according to the data analysis results;

[0020] The deep learning algorithm is used to enhance the diversified NMR sample data or generate new NMR sample data, thereby completing the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report.

[0021] In a specific implementation scheme, the steps of completing the standardization of the NMR spectrum data and the NMR examination report are specifically:

[0022] Match the enhanced or newly generated NMR sample data with the NMR inspection report to obtain the matching key parameters;

[0023] According to the key parameters of the matching, a process for standardizing the NMR spectrum data and the nuclear magnetic resonance inspection report is formulated to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0024] In a specific implementation scheme, the process of standardizing the NMR spectrum data and nuclear magnetic resonance inspection reports includes specific steps and requirements for each link of data acquisition, preprocessing, analysis, and matching.

[0025] In a specific embodiment, it also includes:

[0026] Establish a standardized database.

[0027] In a specific implementation scheme, the steps of establishing a standardized database are as follows:

[0028] A standardized NMR spectrum database and a nuclear magnetic resonance inspection report database are established to store and manage the analyzed NMR sample data and nuclear magnetic resonance inspection reports according to unified standards.

[0029] In the second aspect, the present application provides a digital auxiliary device for nuclear magnetic resonance examination, including a data acquisition module, a preprocessing module, an AI large model analysis module and a data enhancement module, wherein:

[0030] The data acquisition module is used to obtain NMR spectrum data;

[0031] The preprocessing module is used to preprocess the NMR spectrum data to obtain preprocessed NMR spectrum data;

[0032] The AI ​​large model parsing module is used to automatically parse the preprocessed NMR spectrum data using the pretrained AI large model to obtain parsed NMR spectrum data;

[0033] The data enhancement module is used to generate diversified NMR sample data based on the analyzed NMR spectrum data, and match the NMR sample data with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0034] In the above technical scheme, NMR spectrum data is acquired; the NMR spectrum data is preprocessed to obtain preprocessed NMR spectrum data; the preprocessed NMR spectrum data is automatically parsed using a pretrained AI large model to obtain parsed NMR spectrum data; diversified NMR sample data is generated based on the parsed NMR spectrum data, and the NMR sample data is matched with the nuclear magnetic resonance examination report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report; the analysis efficiency and accuracy are improved, and data enhancement is performed at the same time to improve the accuracy, standardization and reliability of nuclear magnetic resonance examination; the demand and dependence on professional doctors are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1A flowchart of the NMR standardization method provided in the embodiment of the present application;

[0036] Figure 2 This is a structural block diagram of the digital auxiliary device for nuclear magnetic resonance examination provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] The present application is further described in detail below through the accompanying drawings and embodiments. Through these descriptions, the characteristics and advantages of the present application will become clearer and more specific.

[0038] The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise noted.

[0039] In addition, the technical features involved in different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0040] To facilitate understanding of the nuclear magnetic examination digital auxiliary device provided in the embodiment of the present application, its application scenario is first explained. The nuclear magnetic examination digital auxiliary device provided in the embodiment of the present application is used to automatically analyze NMR data through an AI large model, improve the analysis efficiency and accuracy, and perform data enhancement at the same time to improve the accuracy, standardization and reliability of nuclear magnetic examination. With the improvement of public health awareness and the aging of the population, the demand for nuclear magnetic examination is gradually increasing, and the number of nuclear magnetic doctors is extremely short. This causes nuclear magnetic doctors to often work overloaded, and patients need to wait for a period of time after nuclear magnetic examination to get the report results. If the level of nuclear magnetic doctors is not good, misdiagnosis or missed diagnosis may occur. In addition, excessive workload may also cause fatigue of nuclear magnetic doctors and misdiagnosis or missed diagnosis, which will bring a series of adverse consequences to the clinical treatment of patients. For this reason, the embodiment of the present application provides a nuclear magnetic examination digital auxiliary device to automatically analyze NMR data through an AI large model, improve the analysis efficiency and accuracy, and perform data enhancement at the same time to improve the accuracy, standardization and reliability of nuclear magnetic examination. The following is a detailed description of the embodiment in conjunction with specific drawings.

[0041] refer to Figure 1 and Figure 2 , Figure 1 A reference diagram of the usage status of the digital auxiliary device for nuclear magnetic resonance examination provided in an embodiment of the present application is shown.

[0042] exist Figure 1 In the present application, an embodiment provides a nuclear magnetic resonance standardization method, comprising the following steps:

[0043] Obtain NMR spectrum data;

[0044] Preprocessing the NMR spectrum data to obtain preprocessed NMR spectrum data;

[0045] Automatically parsing the preprocessed NMR spectrum data using a pretrained AI large model to obtain parsed NMR spectrum data;

[0046] Based on the analyzed NMR spectrum data, diversified NMR sample data are generated, and the NMR sample data are matched with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0047] In the above technical scheme, NMR spectrum data is acquired; the NMR spectrum data is preprocessed to obtain preprocessed NMR spectrum data; the preprocessed NMR spectrum data is automatically parsed using a pretrained AI large model to obtain parsed NMR spectrum data; diversified NMR sample data is generated based on the parsed NMR spectrum data, and the NMR sample data is matched with the nuclear magnetic resonance examination report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report; the analysis efficiency and accuracy are improved, and data enhancement is performed at the same time to improve the accuracy, standardization and reliability of nuclear magnetic resonance examination; the demand and dependence on professional doctors are reduced.

[0048] The digital auxiliary device for nuclear magnetic resonance examination in the embodiment of the present application is like an excellent and tireless nuclear magnetic resonance doctor, which can work 24 hours a day, and can ensure that the patient's imaging diagnosis needs can be handled in a timely manner even on holidays, thus greatly alleviating the shortage of nuclear magnetic resonance doctors and the problem of excessive workload, thereby improving the diagnostic efficiency while also improving the accuracy of diagnosis, providing patients with faster and better services, thereby effectively improving patient satisfaction. In addition, in areas where medical conditions are not yet perfect, the device can achieve expert-level imaging diagnosis, which will greatly improve the level of medical services.

[0049] In addition, the digital auxiliary device for MRI examination has a learning function, which can realize the continuous digital storage of MRI examination films and the storage of professional pictures, so that it can gradually mature and eventually form an expert-level digital MRI doctor.

[0050] In a specific implementation scheme, the step of preprocessing the NMR spectrum data to obtain the preprocessed NMR spectrum data is specifically:

[0051] The NMR spectrum data is cleaned and denoised to obtain pre-processed NMR spectrum data.

[0052] In a specific implementation scheme, the pre-processed NMR spectrum data is automatically parsed using a pre-trained AI large model to obtain the steps of parsing the NMR spectrum data, specifically:

[0053] First, the NMR spectra and their corresponding molecular structure data are used for training to learn the mapping relationship between the NMR spectra and the molecular structure;

[0054] During the analysis process, the pre-processed NMR spectrum data is input, and the corresponding analyzed NMR spectrum data is automatically output using the AI ​​large model.

[0055] In a specific embodiment, the analyzed NMR spectrum data includes molecular structures and their probability rankings.

[0056] In a specific implementation scheme, the steps of generating diversified NMR sample data based on the analyzed NMR spectrum data, matching the NMR sample data with the nuclear magnetic resonance examination report, and completing the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report are specifically as follows:

[0057] Based on the analyzed NMR spectrum data, each peak in the spectrum is identified and its position and intensity are determined;

[0058] Use fitting methods to process overlapping peaks or peaks with complex shapes and obtain peak parameters as data analysis results;

[0059] Generating diversified NMR sample data according to the data analysis results;

[0060] The deep learning algorithm is used to enhance the diversified NMR sample data or generate new NMR sample data, thereby completing the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report.

[0061] In a specific implementation scheme, the steps of completing the standardization of the NMR spectrum data and the NMR examination report are specifically:

[0062] Match the enhanced or newly generated NMR sample data with the NMR inspection report to obtain the matching key parameters;

[0063] According to the key parameters of the matching, a process for standardizing the NMR spectrum data and the nuclear magnetic resonance inspection report is formulated to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0064] In a specific implementation scheme, the process of standardizing the NMR spectrum data and nuclear magnetic resonance inspection reports includes specific steps and requirements for each link of data acquisition, preprocessing, analysis, and matching.

[0065] In a specific embodiment, it also includes:

[0066] Establish a standardized database.

[0067] In a specific implementation scheme, the steps of establishing a standardized database are as follows:

[0068] A standardized NMR spectrum database and a nuclear magnetic resonance inspection report database are established to store and manage the analyzed NMR sample data and nuclear magnetic resonance inspection reports according to unified standards.

[0069] exist Figure 2 In the embodiment of the present application, a digital auxiliary device for nuclear magnetic resonance examination is also provided, including a data acquisition module, a preprocessing module, an AI large model analysis module and a data enhancement module, wherein:

[0070] The data acquisition module is used to obtain NMR spectrum data;

[0071] The preprocessing module is used to preprocess the NMR spectrum data to obtain preprocessed NMR spectrum data;

[0072] The AI ​​large model parsing module is used to automatically parse the preprocessed NMR spectrum data using the pretrained AI large model to obtain parsed NMR spectrum data;

[0073] The data enhancement module is used to generate diversified NMR sample data based on the analyzed NMR spectrum data, and match the NMR sample data with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

[0074] In the above technical scheme, NMR spectrum data is acquired; the NMR spectrum data is preprocessed to obtain preprocessed NMR spectrum data; the preprocessed NMR spectrum data is automatically parsed using a pretrained AI large model to obtain parsed NMR spectrum data; diversified NMR sample data is generated based on the parsed NMR spectrum data, and the NMR sample data is matched with the nuclear magnetic resonance examination report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report; the analysis efficiency and accuracy are improved, and data enhancement is performed at the same time to improve the accuracy, standardization and reliability of nuclear magnetic resonance examination; the demand and dependence on professional doctors are reduced.

[0075] It should be noted that the AI ​​large model analysis module and the data enhancement module are implemented through a processor or a server; the data acquisition module is implemented through a nuclear magnetic resonance device.

[0076] AI big models refer to artificial intelligence models with a large number of parameters (usually ranging from millions to billions or even trillions) and complex structures. These models can learn and extract complex patterns and laws from large-scale data through deep learning technology, thereby demonstrating excellent performance in various tasks.

[0077] The AI ​​big model has the following characteristics:

[0078] 1. Large-scale parameter volume: The core of large AI models lies in their large parameter volume, which is used to store the information and knowledge learned by the model, enabling it to represent complex language structures, image features, or other types of data features;

[0079] 2. Complex structure: AI large models usually use deep neural network structures, such as the Transformer architecture. These network structures can extract high-order features from data and establish complex relationships between features through multi-level nonlinear transformations and activation functions;

[0080] 3. High performance: Due to their large number of parameters and complex structures, large AI models perform well on various tasks, such as natural language processing, computer vision, speech recognition, etc.

[0081] 4. Transfer learning capability: AI large models can be pre-trained in one or more fields and can transfer the learned knowledge to new tasks or new fields. This transfer learning capability enables the model to learn new tasks faster and also improves the performance of the model in unknown fields.

[0082] The existing major AI models occupy an important position in the field of artificial intelligence, mainly including:

[0083] 1. OpenAI series

[0084] GPT Series

[0085] GPT-1: A pre-trained language model released by OpenAI in 2018, which opened a new chapter in natural language processing.

[0086] GPT-2: Further enhanced the model size and capabilities, but the full model was not made public to avoid abuse.

[0087] GPT-3: released in 2020, with 175 billion parameters, it demonstrates powerful text generation and comprehension capabilities.

[0088] GPT-4: released in 2023, supports multimodal understanding and content generation, including text, images, and videos.

[0089] Sora: Released in 2024, it focuses on video generation and proposes spatiotemporal fragmentation and diffusion Transformer technologies.

[0090] ChatGPT: A chatbot model based on the GPT architecture, which has powerful natural language interaction and generation capabilities and is widely used in dialogue systems and customer service fields.

[0091] 2. Google Series

[0092] BERT: A pre-trained language model released by Google in 2018. It is as famous as GPT and improves the natural language processing effect through pre-training tasks.

[0093] Multilingual BERT: A BERT model that supports multiple languages ​​and enhances cross-language processing capabilities.

[0094] GEMINI: Google's large multimodal model that supports conversion recognition between image, audio, and video modalities and can handle complex multimodal tasks.

[0095] GEMMA: Google's open source series of models, with 2 billion and 7 billion parameter versions, which perform well on multiple benchmark tests.

[0096] 3. Meta (formerly Facebook) series

[0097] LLaMA: A series of large models developed by Meta, including versions of different sizes, such as 8 billion parameters and 70 billion parameters, and a large version with more than 400 billion parameters. LLaMA performed well in the MMLU test, showing strong language understanding capabilities.

[0098] 4. Major domestic models

[0099] GLM-4: A large language model developed in China. In the evaluation, it showed performance close to that of the world's top models, especially in Chinese scenarios.

[0100] Wenxinyiyan 4.0: An industrial-grade knowledge-enhanced large model launched by Baidu, which achieves efficient content generation and comprehension capabilities through the integrated learning of massive data and large-scale knowledge.

[0101] Tongyi Qianwen: Alibaba's big model, based on a unified base, has built a hierarchical model system that covers multiple fields such as natural language processing, multimodality, and computer vision.

[0102] Pangu Big Model: The big model launched by Huawei Cloud emphasizes three core designs: model data, network structure, and generalization capability, and supports applications in multiple industries and segmented scenarios.

[0103] Hunyuan Big Model: Tencent's big model covers NLP, CV, multimodal big models, etc., demonstrating its leading position in multiple authoritative data set lists.

[0104] Those skilled in the art will appreciate that the present application may be implemented as a system, method or computer program product.

[0105] Therefore, the present disclosure may be specifically implemented in the following forms, namely: it may be completely hardware, it may be completely software (including firmware, resident software, microcode, etc.), or it may be a combination of hardware and software, generally referred to herein as a "circuit", "module" or "system". In addition, in some embodiments, the present application may also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium may contain computer-readable program code.

[0106] Any combination of one or more computer-readable media can be used. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this document, computer-readable storage media can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0107] Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art can change, modify, replace and modify the above embodiments within the scope of the present application. On this basis, a variety of replacements and improvements can be made to the present application, all of which fall within the scope of protection of the present application.

Claims

1. A nuclear magnetic standardization method, characterized in that The following steps are involved: Obtain NMR spectrum data; Preprocessing the NMR spectrum data to obtain preprocessed NMR spectrum data; Automatically parsing the preprocessed NMR spectrum data using a pretrained AI large model to obtain parsed NMR spectrum data; Based on the analyzed NMR spectrum data, diversified NMR sample data are generated, and the NMR sample data are matched with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

2. The nuclear magnetic standardization method according to claim 1, characterized in that The step of preprocessing the NMR spectrum data to obtain the preprocessed NMR spectrum data is specifically: The NMR spectrum data is cleaned and denoised to obtain pre-processed NMR spectrum data.

3. The nuclear magnetic standardization method according to claim 2, characterized in that The pre-processed NMR spectrum data is automatically parsed using a pre-trained AI large model to obtain the steps of parsing the NMR spectrum data, specifically: First, the NMR spectra and their corresponding molecular structure data are used for training to learn the mapping relationship between the NMR spectra and the molecular structure; During the analysis process, the pre-processed NMR spectrum data is input, and the corresponding analyzed NMR spectrum data is automatically output using the AI ​​large model.

4. The nuclear magnetic standardization method according to claim 3, characterized in that The analyzed NMR spectrum data includes molecular structures and their probability rankings.

5. The nuclear magnetic standardization method according to claim 4, characterized in that The steps of generating diversified NMR sample data based on the analyzed NMR spectrum data, matching the NMR sample data with the nuclear magnetic resonance inspection report, and completing the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report are specifically as follows: Based on the analyzed NMR spectrum data, each peak in the spectrum is identified and its position and intensity are determined; Use fitting methods to process overlapping peaks or peaks with complex shapes and obtain peak parameters as data analysis results; Generating diversified NMR sample data according to the data analysis results; The deep learning algorithm is used to enhance the diversified NMR sample data or generate new NMR sample data, thereby completing the standardization of the NMR spectrum data and the nuclear magnetic resonance examination report.

6. The nuclear magnetic standardization method according to claim 5, characterized in that The steps of completing the standardization of the NMR spectrum data and the NMR inspection report are specifically: Match the enhanced or newly generated NMR sample data with the NMR inspection report to obtain the matching key parameters; According to the key parameters of the matching, a process for standardizing the NMR spectrum data and the nuclear magnetic resonance inspection report is formulated to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

7. The nuclear magnetic standardization method according to claim 6, characterized in that The process of standardizing NMR spectrum data and nuclear magnetic resonance inspection reports includes specific steps and requirements for each link of data acquisition, preprocessing, analysis, and matching.

8. The nuclear magnetic standardization method according to claim 7, characterized in that Also includes: Establish a standardized database.

9. The nuclear magnetic standardization method according to claim 8, characterized in that The steps to establish a standardized database are as follows: A standardized NMR spectrum database and a nuclear magnetic resonance inspection report database are established to store and manage the analyzed NMR sample data and nuclear magnetic resonance inspection reports according to unified standards.

10. A digital auxiliary device for nuclear magnetic resonance examination, characterized in that: It includes data acquisition module, preprocessing module, AI large model analysis module and data enhancement module, among which, The data acquisition module is used to obtain NMR spectrum data; The preprocessing module is used to preprocess the NMR spectrum data to obtain preprocessed NMR spectrum data; The AI ​​large model parsing module is used to automatically parse the preprocessed NMR spectrum data using the pretrained AI large model to obtain parsed NMR spectrum data; The data enhancement module is used to generate diversified NMR sample data based on the analyzed NMR spectrum data, and match the NMR sample data with the nuclear magnetic resonance inspection report to complete the standardization of the NMR spectrum data and the nuclear magnetic resonance inspection report.

Images