DICOM image sequence splitting system of nuclear magnetic resonance

Abstract

The invention discloses a nuclear magnetic resonance DICOM image sequence splitting system, which comprises an image transmission module, an image reading and analyzing module, a sequence splitting and analyzing module, a data storage module and an image display module, and is characterized in that the image transmission module, the image reading and analyzing module, the sequence splitting and analyzing module, the data storage module and the image display module are in communication connection; a fully automatic process is realized from image receiving, analysis, splitting to display. The invention provides an image processing platform which is efficient, accurate, flexible and easy to be integrated into the existing medical workflow so as to better serve clinical practice and scientific research.

Description

Technical Field

[0001] The present invention belongs to the technical field of commercial vehicle driving safety and management, and in particular relates to a DICOM image sequence splitting system for nuclear magnetic resonance. Background Art

[0002] In modern medicine, medical imaging (such as X-rays, CT scans, and MRIs) is crucial for disease diagnosis and treatment planning. Magnetic resonance imaging (MRI), in particular, provides high-resolution soft tissue contrast images and is widely used in fields such as neuroscience, oncology, and cardiology. However, with advances in medical technology and improvements in imaging equipment performance, the amount of image data generated has increased dramatically, making effective management and analysis of this data increasingly challenging.

[0003] To ensure interoperability and data exchange between medical imaging devices from different manufacturers, the international DICOM (Digital Imaging and Communications in Medicine) standard was established. This standard not only specifies the image file format but also defines a series of metadata tags that describe important information such as image acquisition conditions, patient information, and scanning parameters. These tags provide the foundation for image storage, transmission, display, and further analysis and processing.

[0004] Magnetic resonance imaging typically consists of a series of two- or three-dimensional images acquired according to a specific protocol. Each sequence may reflect different physical characteristics, such as T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI). In particular, the B value in diffusion-weighted imaging (corresponding to the DICOM tag (0018,9087)) is a measure of the diffusion of water molecules and is particularly important for the research and diagnosis of diseases such as stroke.

[0005] Traditionally, physicians have relied on their experience and expertise to identify and categorize these image sequences. This process is time-consuming, labor-intensive, and subject to subjective factors. Furthermore, the varying brands and models of MRI equipment used by different medical institutions result in differences in image formats and labeling, complicating cross-institutional image sharing and comparison. While some advanced PACS (Picture Archiving and Communication Systems) systems have begun to offer rudimentary image sorting capabilities, they often require manual intervention to ensure accurate classification and fall short of full automation.

[0006] With the growth of medical imaging data volume and the improvement of technical complexity, the urgent need for image processing automation is becoming increasingly apparent.

[0007] In the field of MRI DICOM image processing, existing technical solutions mainly include: 1. Rule-based image classification: This type of system automatically identifies and classifies image sequences through a set of preset rules or templates, and assigns images to different categories; 2. Semi-automated workflow: Some advanced PACS systems provide preliminary image sorting functions, but the final confirmation or adjustment still requires user participation. These systems may perform preliminary classification of images based on DICOM metadata, which is then reviewed and revised by doctors or technicians; 3. Image analysis based on machine learning: In recent years, machine learning algorithms have been increasingly used in medical image analysis. These algorithms can learn from a large number of labeled image samples to improve classification accuracy. In particular, deep learning models have shown strong capabilities in image recognition tasks. These methods attempt to solve the problem of automated segmentation and management of image sequences to varying degrees, but each has its own scope of application and limitations: (1) High dependence on manual processing: Traditional methods rely heavily on doctors' experience and naked eye recognition to segment and classify image sequences. This is not only time-consuming and labor-intensive, but also easily influenced by subjective factors, leading to misjudgments or omissions. (2) Insufficient flexibility of rule systems: Although rule-based classification systems are simple and direct, they have poor flexibility and adaptability when faced with non-standard naming, new types of scans, or inaccurate DICOM tags, making it difficult to maintain sustained accuracy. (3) Limited efficiency of semi-automated processes: The preliminary sorting functions provided by some advanced PACS systems still require manual review and adjustment, and have not been fully automated, which limits the improvement of work efficiency. In addition, there may be differences in judgment between different users, affecting the consistency of results. (4) Insufficient specificity of machine learning models: Although machine learning algorithms perform well in image recognition, general models are difficult to directly apply to MRI-specific parameters (such as B value). In addition, training these models requires a large amount of high-quality annotated data, which is challenging to obtain in the medical field. At the same time, the decision-making process of complex models is often difficult to explain, which places higher demands on medical applications. (5) Cross-institutional compatibility and standardization issues: Due to the different brands and models of equipment used by different medical institutions, the image formats and labels generated are different, which increases the difficulty of cross-institutional image sharing and comparison and hinders the standardization process. Summary of the Invention

[0008] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a DICOM image sequence splitting system for nuclear magnetic resonance imaging to solve the problem of low processing efficiency and accuracy of DICOM image sequences for nuclear magnetic resonance imaging.

[0009] To solve the above technical problems, the present invention is implemented by adopting the following solutions: The present invention provides a DICOM image sequence splitting system for nuclear magnetic resonance imaging, comprising: The image transmission module is used to receive the image sequence sent by the PACS through the DICOM network transmission protocol and send the image sequence to the image reading and parsing module; Image reading and parsing module, used to read the image files of the image sequence according to the DICOM standard protocol and parse the metadata in the image files; The sequence splitting analysis module is used to extract data related to sequence splitting from metadata and analyze the extracted data to obtain the results of image sequence splitting dimensions; A data storage module is used to store image files in a file server, store core metadata in the image files in a database, and define the result of splitting the image sequence dimensions as a split identifier and record it in the database; Image display module, including: The client is used to obtain and decide whether to call the corresponding interface of the server based on the split identifier of the image sequence in the database; obtain the image list under the image sequence and the image files under the image list from the server through the UID of the image sequence; The server is used to respond to the client's interface call instructions, and concatenate the UID of the image sequence with the corresponding split identifier and split value as the UID of the new image sequence and return it to the client; based on the UID of the image sequence sent by the client and query the core metadata in the database to obtain the image list under the image sequence and return it to the client, and access the corresponding image file in the file server and return it to the client.

[0010] Optionally, extract metadata related to sequence splitting, including: Check the device type, the corresponding DICOM tag is (0008,0060), which indicates the type of device used for imaging sequence; Extract the B value, the corresponding DICOM tag is (0018,9087), which represents the measurement value of the gradient applied in the image sequence; Extract the scan time stamp, which corresponds to the DICOM tag (0020,0100), indicating the time sequence of the dynamic or functional image set; Extract the spatial position data. The corresponding DICOM tag is (0020, 0032), which indicates the starting position of the image.

[0011] Optionally, the extracted data is analyzed to obtain the image sequence splitting dimension results, including: Determine whether the device type is MRI. If not, the image sequence will not be split. If so, determine whether each image in the image sequence has a B value and whether the B value is partially repeated. If so, the image sequence will be split according to the B value. If not, determine whether each image in the image sequence has a scan time identifier and whether the scan time identifier value is partially repeated. If so, the image sequence will be split according to the scan time identifier. If not, determine whether each image in the image sequence has spatial position data and whether the spatial position data is partially repeated. If so, the image sequence will be split according to the spatial position data. If not, the image sequence will not be split.

[0012] Optionally, the result of splitting the image sequence dimension is defined as a splitting identifier, including: defining 0 to indicate no splitting, 1 to indicate splitting by B value, 2 to indicate splitting by scan time identifier, and 3 to indicate splitting by spatial position data.

[0013] Optionally, the database storage structure is stored in four dimensions, namely, patient, examination, sequence, and image instance, according to the DICOM model. The result of splitting the image sequence dimension is defined as a split identifier recorded in the sequence dimension table.

[0014] Optionally, obtaining the split identifier of the image sequence in the database includes: putting the split identifier of the image sequence into the return data of the Series list interface of QIDO-RS through the custom tag of DICOM and returning it to the client.

[0015] Optionally, in response to the client's interface call instruction, the UID of the image sequence is concatenated with the corresponding split identifier and split value and returned to the client as the UID of the new image sequence, including: Query the split identifier of the image sequence according to the original image sequence UID, query the split value list under the image sequence according to different split identifiers, and use the original image sequence UID, the corresponding split identifier and split value to concatenate as the new image sequence UID and return it to the client; Among them, the splicing format of the original image sequence UID and the corresponding split identifier and split value is: original image sequence UID_split identifier_split value.

[0016] Optionally, according to the UID of the image sequence sent by the client, the core metadata in the database is queried to obtain the image list under the image sequence and return it to the client, and the corresponding image file in the file server is accessed and returned to the client, including: The server uses the UID of the image sequence in the Instances interface of QIDO-RS and queries the core metadata in the database to obtain the image list under the image sequence and returns it to the client; If the UID of the image sequence does not carry a split identifier and a split value, all image files under the image sequence in the access file server are returned to the client; otherwise, the image files of the corresponding group under the image sequence in the access file server are returned to the client. Beneficial effects

[0017] The present invention realizes a fully automated process through an image transmission module, an image reading and parsing module, a sequence splitting and analysis module, a data storage module, and an image display module: from image reception, parsing, splitting to display, the entire process requires almost no human intervention, significantly reducing the workload of doctors.

[0018] The multi-dimensional data analysis of the present invention not only considers conventional DICOM tags, but also pays special attention to MRI-specific parameters (such as B value) to achieve more precise sequence segmentation.

[0019] The present invention has wide compatibility: the system design can adapt to different types of MRI scans and equipment, has good scalability, and supports the continuous development of medical imaging technology.

[0020] The present invention has an efficient database design: it adopts a relational database or a NoSQL database, optimizes the storage structure to improve query efficiency, and ensures fast retrieval and access to large amounts of image data.

[0021] The present invention is cross-institutional collaboration: strictly following the DICOM standard, promoting image sharing and communication between different medical institutions, and promoting the standardization process. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is an architecture diagram of a DICOM image sequence splitting system for nuclear magnetic resonance imaging provided by Example 1 of the present invention; Figure 2 This is an interface flow chart for obtaining subsequences after sequence splitting by a server in an image display module of a DICOM image sequence splitting system for nuclear magnetic resonance provided by Example 2 of the present invention; Figure 3 This is a flow chart of analyzing sequence splitting dimensions by a sequence splitting analysis module in a nuclear magnetic resonance DICOM image sequence splitting system provided by Example 2 of the present invention. DETAILED DESCRIPTION

[0023] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. Example 1

[0024] like Figure 1As shown, this embodiment provides a DICOM image sequence splitting system for nuclear magnetic resonance, including an image transmission module, an image reading and parsing module, a sequence splitting and analysis module, a data storage module and an image display module. The image transmission module, the image reading and parsing module, the sequence splitting and analysis module, the data storage module and the image display module are communicated with each other, and a fully automated process is realized from image reception, parsing, splitting to display. The entire process requires almost no human intervention, which significantly reduces the workload of doctors. Example 2

[0025] This embodiment provides a method for splitting a DICOM image sequence of an MRI, which is implemented by the DICOM image sequence splitting system of the MRI described in Example 1. The detailed working steps of each module are as follows: S01, PACS (or imaging equipment) and the system of the present invention are interconnected via the DICOM network transmission protocol; S02. The image transmission module receives the image sequence sent by the PACS system (or imaging device) according to the DICOM network transmission protocol, and then sends it to the image reading and analysis module; S03. The image reading and parsing module reads the image files of the sequence according to the DICOM standard protocol and parses the metadata and pixel data parts therein; S04. The sequence splitting analysis module extracts data related to sequence splitting from the metadata: First, the inspection equipment type, the corresponding DICOM tag is (0008, 0060), which indicates the type of equipment used for imaging the sequence; second, the B value, the corresponding DICOM tag is (0018, 9087), which is a sensitivity to the diffusion movement performance, an indicator for detecting the diffusion movement ability, and a measure of the size of the gradient applied in the sequence; third, the scan time identifier, the corresponding DICOM tag is (0020, 0100), which is used to indicate the time sequence of the dynamic or functional image set; fourth, the spatial position data, the corresponding DICOM tag is (0020, 0032), which indicates the starting position of the image.

[0026] S05, the sequence splitting and analysis module uses the data obtained in step S04 to Figure 3The process analysis shown in the results is to determine whether the sequence needs to be split and, if so, by what dimension. Specifically, first determine whether the device type is MRI (nuclear magnetic resonance imaging). If not, the sequence is not split; if so, then determine whether each image in the sequence has a B value and whether the B value is partially repeated (they have the same B value, but not all the same). If so, the sequence is split by the B value; if not, then determine whether each image in the sequence has a scan time stamp and whether the scan time stamp value is partially repeated (they have the same scan time stamp, but not all the same). If so, the sequence is split by the scan time stamp; if not, then determine whether each image in the sequence has spatial position data and whether the spatial position data is partially repeated (they have the same spatial position data, but not all the same). If so, the sequence is split by the spatial position data; if not, the sequence is not split.

[0027] S06. The data storage module stores the image file in the file server; some core metadata obtained by parsing the image file is stored in the database for easy retrieval. The database storage structure is divided into four dimensions according to the DICOM model, namely: patient (Patient), examination (Study), sequence (Series), and image instance (Instance). The result of the sequence splitting dimension obtained in step S05 is recorded in the sequence (Series) dimension table, which is defined as the splitting identifier: 0 means no splitting, 1 means splitting by B value, 2 means splitting by scan time identifier, and 3 means splitting by spatial position data.

[0028] S07, the image display module is divided into client and server, and is developed in accordance with the DICOM Web protocol. The Series list interface of QIDO-RS returns the series list contained in the study. According to the image sequence splitting identifier stored in step S06, we can put the image sequence splitting identifier into the return data of the Series list interface through the DICOM custom tag and return it to the client. In this way, the client can obtain the splitting identifier of each image sequence and decide whether to call the "Interface for obtaining the subsequences after the image sequence is split" on the server according to this identifier. The workflow of this interface is detailed in the attached Figure 2Specifically, first query the "split identifier" of the image sequence based on the original image sequence UID, then query the "split value" list under the image sequence based on different split identifiers (de-duplication), and finally use the original image sequence UID to underline the "split identifier" and "split value" (i.e.: original image sequence UID_split identifier_split value) as a new image sequence UID list and return it to the client. In this way, when the client later uses the QIDO-RS Instances interface to obtain the Instance list under a series, the "split identifier" and "split value" will be automatically included in the image sequence UID for the split subsequences. The server uses the image sequence UID and queries the core metadata in the database in this interface to obtain the image file instance list under the image sequence and return it to the client. The image file instances under the image file instance list need to access the corresponding image file instances in the file server and return them to the client: if the image sequence UID does not contain the "split identifier" and "split value", all image file instances under the image sequence are obtained; otherwise, the image file instances of the corresponding group under the image sequence are obtained.

[0029] In summary, the present invention has the following advantages: 1. Multi-dimensional data analysis: not only conventional DICOM tags are considered, but also special attention is paid to MRI-specific parameters (such as B-value) to achieve more refined sequence splitting; 2. Wide compatibility: The system design can adapt to different types of MRI scans and equipment, has good scalability, and supports the continuous development of medical imaging technology; 3. Efficient database design: Adopting relational databases or NoSQL databases, optimizing storage structures to improve query efficiency, and ensuring rapid retrieval and access to large amounts of image data; 4. Cross-institutional collaboration: Strictly adhere to the DICOM standard, promote image sharing and communication between different medical institutions, and promote the standardization process.

[0030] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A DICOM image sequence splitting system for nuclear magnetic resonance imaging, characterized in that: include: The image transmission module is used to receive the image sequence sent by the PACS through the DICOM network transmission protocol and send the image sequence to the image reading and parsing module; Image reading and parsing module, used to read the image files of the image sequence according to the DICOM standard protocol and parse the metadata in the image files; The sequence splitting analysis module is used to extract data related to sequence splitting from metadata and analyze the extracted data to obtain the results of image sequence splitting dimensions; A data storage module is used to store image files in a file server, store core metadata in the image files in a database, and define the result of splitting the image sequence dimensions as a split identifier and record it in the database; Image display module, including: The client is used to obtain and decide whether to call the corresponding interface of the server based on the split identifier of the image sequence in the database; Obtain the image list and image files under the image list from the server through the UID of the image sequence; The server is used to respond to the client's interface call instructions, and concatenate the UID of the image sequence with the corresponding split identifier and split value as the UID of the new image sequence and return it to the client; based on the UID of the image sequence sent by the client and query the core metadata in the database to obtain the image list under the image sequence and return it to the client, and access the corresponding image file in the file server and return it to the client.

2. The DICOM image sequence splitting system for nuclear magnetic resonance according to claim 1, characterized in that: Extract metadata related to sequence splitting, including: Check the device type, the corresponding DICOM tag is (0008,0060), which indicates the type of device used for imaging sequence; Extract the B value, the corresponding DICOM tag is (0018,9087), which represents the measurement value of the gradient applied in the image sequence; Extract the scan time stamp, which corresponds to the DICOM tag (0020,0100), indicating the time sequence of the dynamic or functional image set; Extract the spatial position data. The corresponding DICOM tag is (0020, 0032), which indicates the starting position of the image.

3. The DICOM image sequence segmentation system for nuclear magnetic resonance according to claim 1, characterized in that: The extracted data is analyzed to obtain the image sequence splitting dimension results, including: Determine whether the device type is MRI. If not, the image sequence will not be split. If so, determine whether each image in the image sequence has a B value and whether the B value is partially repeated. If so, the image sequence will be split according to the B value. If not, determine whether each image in the image sequence has a scan time identifier and whether the scan time identifier value is partially repeated. If so, the image sequence will be split according to the scan time identifier. If not, determine whether each image in the image sequence has spatial position data and whether the spatial position data is partially repeated. If so, the image sequence will be split according to the spatial position data. If not, the image sequence will not be split.

4. The DICOM image sequence splitting system for nuclear magnetic resonance according to claim 1, characterized in that: The result of splitting the image sequence dimension is defined as the splitting identifier, including: 0 means no splitting, 1 means splitting by B value, 2 means splitting by scanning time identifier, and 3 means splitting by spatial position data.

5. The DICOM image sequence splitting system for nuclear magnetic resonance according to claim 1, characterized in that: The database storage structure is divided into four dimensions according to the DICOM model: patient, examination, sequence, and image instance. The result of the image sequence split dimension is defined as the split identifier and recorded in the sequence dimension table.

6. The DICOM image sequence splitting system for nuclear magnetic resonance according to claim 1, characterized in that: Obtain the split identifier of the image sequence in the database, including: putting the split identifier of the image sequence into the return data of the Series list interface of QIDO-RS through the custom tag of DICOM and returning it to the client.

7. The DICOM image sequence segmentation system for nuclear magnetic resonance according to claim 1, characterized in that: In response to the client's interface call instruction, the UID of the image sequence is concatenated with the corresponding split identifier and split value as the UID of the new image sequence and returned to the client, including: Query the split identifier of the image sequence according to the original image sequence UID, query the split value list under the image sequence according to different split identifiers, and use the original image sequence UID, the corresponding split identifier and split value to concatenate as the new image sequence UID and return it to the client; Among them, the splicing format of the original image sequence UID and the corresponding split identifier and split value is: original image sequence UID_split identifier_split value.

8. The DICOM image sequence splitting system for nuclear magnetic resonance according to claim 1, characterized in that: Based on the UID of the image sequence sent by the client, the core metadata in the database is queried to obtain the image list under the image sequence and return it to the client. The corresponding image file in the file server is accessed and returned to the client, including: The server uses the UID of the image sequence in the Instances interface of QIDO-RS and queries the core metadata in the database to obtain the image list under the image sequence and returns it to the client; If the UID of the image sequence does not carry a split identifier and a split value, all image files under the image sequence in the access file server are returned to the client; otherwise, the image files of the corresponding group under the image sequence in the access file server are returned to the client.

Images