Medical image big data platform desensitization and lossless compression processing method and device in multi-modal data acquisition process and electronic device

Abstract

This application discloses a method, apparatus, and electronic device for desensitization and lossless compression processing during multimodal data acquisition in a medical imaging big data platform. The method includes: in a DICOM C-STORE SCP receiving link, before the DICOM file is stored, performing real-time preprocessing on the received DICOM data stream; parsing all original tag sets within the data stream to generate first data containing complete original plaintext metadata; desensitizing the DICOM tags based on the first data to generate second data with a desensitized tag set; and finally, compressing the image pixel data in the data stream based on the second data to output the processed DICOM file. This solution achieves real-time online preprocessing of the DICOM data stream, effectively protecting patient privacy information while reducing image data volume, significantly improving medical image transmission efficiency and system storage performance.

Description

Technical Field

[0001] This application relates to the field of medical image data processing technology, and in particular to a method, apparatus and electronic device for desensitization and lossless compression processing of multimodal data acquisition in a medical image big data platform. Background Technology

[0002] With the rapid development of medical informatization, the demand for medical imaging data processing is increasing in various business scenarios such as PACS systems, medical imaging cloud platforms, medical insurance image review, cross-hospital data sharing, and hospital image data collection and archiving. Medical imaging data contains a large amount of patient privacy information and key diagnostic data. In order to protect patient privacy and security, improve data transmission efficiency, and meet medical compliance requirements, effective de-identification and lossless compression processing is required during the medical imaging data collection process.

[0003] Currently, in medical image acquisition and transmission systems, medical imaging equipment (CT, DR, MRI, ultrasound, etc.) all adhere to the DICOM 3.0 standard, uploading raw DICOM image files to the SCP storage service via the C-STORE communication protocol. The industry-standard processing method is as follows: after receiving the complete DICOM file, the SCP server directly caches, forwards, or archives the raw file locally without pre-processing structurally. For desensitization requirements, post-processing batch desensitization or fixed-label desensitization is commonly used. For compression requirements, direct transmission of the raw file or common compression algorithms (JPEG, ZIP compression) are typically employed to complete data processing.

[0004] However, the above processing methods have many problems: low metadata reading efficiency, traditional DICOM file metadata and pixel data are coupled and stored, business queries and data audits must parse the complete file, resulting in high IO overhead; high transmission bandwidth consumption, the original DICOM file is large in size, consuming a lot of bandwidth during transmission, high transmission latency, and network congestion and image packet loss during peak periods; poor compression adaptability, existing compression methods are either lossy compression that destroys diagnostic pixel data, or compression algorithms designed for general files, resulting in low compression ratios for medical grayscale images, and failing to balance lossless and compression efficiency; strong data coupling, metadata and image pixel data cannot be separated and managed, which is not conducive to hierarchical storage of medical data, desensitization processing, and lightweight data distribution; poor desensitization flexibility, original data cannot be retained, traditional desensitization methods have fixed tags, cannot customize sensitive tags, the original plaintext information is lost after desensitization, and it is impossible to balance privacy protection and compliance traceability.

[0005] Therefore, the desensitization and lossless compression methods in the medical image data acquisition process of related technologies have problems such as low metadata reading efficiency, high transmission bandwidth consumption, poor compression adaptability, strong data coupling, poor desensitization flexibility, and the inability to retain original data, resulting in low medical image data processing efficiency, poor security, and insufficient compliance. Summary of the Invention

[0006] The purpose of this application is to provide a method, apparatus, and electronic device for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform, which can process DICOM data streams in real time, effectively protect patient privacy, and significantly improve data transmission efficiency and storage performance.

[0007] In the first aspect, this application discloses a method for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform, the technical solution of which is as follows: In the DICOM C-STORE SCP receiving link, before the DICOM file is stored, the received DICOM data stream is processed in real time; Parse the complete set of raw tags in the DICOM data stream to generate the first data containing all raw plaintext metadata; Based on the first data, the tags in the DICOM data stream are desensitized to generate the second data containing the desensitized tag set. Based on the second data, the pixel data in the DICOM data stream is compressed to generate a processed DICOM file.

[0008] Optionally, the entire set of raw tags in the DICOM data stream is parsed to generate first data containing all raw plaintext metadata, including: Extract all DICOM tags from the DICOM data stream; Serialize all DICOM tags in plaintext JSON format; Generate a separate metadata file to permanently store all original plaintext metadata.

[0009] Optionally, based on the first data, the tags in the DICOM data stream are de-identified to generate second data containing a set of de-identified tags, including: Write the desensitized tag set into the header of the DICOM data stream; A dual metadata architecture is formed, which separates and stores the original plaintext metadata from the de-identified tags; The original plaintext metadata is stored in a separate metadata file, while the de-identified tags are stored inside the DICOM file.

[0010] Optionally, the tags in the DICOM data stream are de-identified, including: Read the pre-configured list of sensitive tags; Iterate through the tags in the DICOM data stream and match them with the list of sensitive tags; Perform a preset replacement operation on the successfully matched sensitive tags.

[0011] Optionally, it also includes: Provides a dynamic configuration interface; The sensitive TAG list can be customized by adding or deleting tags through a dynamic configuration interface; Update the list of sensitive tags and make it effective.

[0012] Optionally, based on the second data, the pixel data in the DICOM data stream is compressed, including: Separate the metadata segment and pixel data segment in the DICOM data stream; The plaintext format of the metadata segment is preserved without compression. Perform JPEG-LS lossless compression on the pixel data segment.

[0013] Optionally, JPEG-LS lossless compression is performed on the pixel data segments, including: Read the raw pixel values ​​from the pixel data segment; The original pixel values ​​are losslessly encoded using the JPEG-LS encoding algorithm; Replace the original pixel data segment in the DICOM data stream with the compressed pixel data segment.

[0014] Optionally, it also includes: After the first data is generated, immediately transmit the metadata file containing all the original plaintext metadata. After generating the processed DICOM file, the compressed and desensitized DICOM image file is then transferred.

[0015] Optionally, a metadata file containing all the original plaintext metadata is transmitted, including: Prioritize transmitting plaintext JSON metadata; The receiving end performs image retrieval and index construction in advance based on JSON plaintext metadata; The system then receives the compressed and desensitized DICOM image files and completes the data association.

[0016] Optionally, during the DICOM C-STORE SCP receiving link, before the DICOM file is stored on disk, the received DICOM data stream is processed in real time, including: Intercept the DICOM data stream in the C-STORE request; Complete all parsing, desensitization, and compression processes in memory; After processing is complete, the processed DICOM file is written to the storage medium.

[0017] Secondly, this application also discloses a device for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform, the technical solution of which is as follows: The receiving and processing unit is used to process the received DICOM data stream in real time during the DICOM C-STORE SCP receiving link, before the DICOM file is stored on the ground. The parsing unit is used to parse the entire set of raw tags in the DICOM data stream and generate the first data containing all raw plaintext metadata. The desensitization processing unit is used to desensitize the tags in the DICOM data stream based on the first data, and generate second data containing the desensitized tag set; The compression processing unit is used to compress the pixel data in the DICOM data stream based on the second data to generate a processed DICOM file.

[0018] Thirdly, this application also discloses a computer-readable storage medium, the technical solution of which is as follows: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the aforementioned desensitization and lossless compression processing methods of the medical imaging big data platform during multimodal data acquisition.

[0019] Fourthly, this application also discloses an electronic device, the technical solution of which is as follows: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by at least one processor. The computer program is executed by at least one processor to enable the at least one processor to perform the desensitization and lossless compression processing method of the above-mentioned medical imaging big data platform in the multimodal data acquisition process.

[0020] Compared to existing technologies, this application provides a method, apparatus, and electronic device for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform. By performing real-time parsing, desensitization, and compression processing on the data stream before the DICOM file is stored, it achieves separate storage and optimized management of metadata and pixel data, effectively avoiding the IO overhead and processing delay caused by traditional coupled storage. At the same time, it takes into account both privacy protection and data integrity, and can process DICOM data streams in real time, effectively protect patient privacy, and significantly improve data transmission efficiency and storage performance. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a method for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform, as provided in an embodiment of this application.

[0023] Figure 2 This is a schematic diagram of the structure of a desensitization and lossless compression processing device for a medical imaging big data platform during multimodal data acquisition, provided in an embodiment of this application.

[0024] Figure 3 This is a schematic diagram of the structural framework of an electronic device provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0026] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] like Figure 1 As shown in this embodiment, a method for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform is presented.

[0028] First, the received DICOM data stream is processed in real time in the DICOM C-STORE SCP receiving link before the DICOM file is stored.

[0029] As one implementation, a middleware service is deployed in the SCP receive link, configured to intercept and process incoming DICOM data streams before the DICOM file is written to the storage medium. This service can be a standalone process or module designed to intervene before the data stream reaches the storage layer, enabling real-time data manipulation. For example, real-time capture of the data stream can be achieved by configuring a network proxy or file system hooks.

[0030] In one embodiment, the entire set of raw tags in the DICOM data stream is parsed to generate first data containing all raw plaintext metadata.

[0031] Specifically, the received DICOM data stream is fed into a parsing module. This parsing module is configured to read the data stream byte-by-byte or block-by-block, identifying and extracting all DICOM tags and their corresponding values. These tags and values ​​are organized into an internal data structure, such as a hash table or tree structure, to represent the original plaintext metadata. This internal data structure is considered the first data, containing the complete set of original tags from the DICOM data stream.

[0032] Based on this, the tags in the DICOM data stream are desensitized using the first data, generating a second data set containing the desensitized tags.

[0033] Specifically, after the first set of data is generated, a de-identification module is activated. This module is configured to access the original plaintext metadata contained in the first set of data. This module is designed to identify specific tags that may contain patient privacy information, such as patient name and ID number. For the identified sensitive tags, their corresponding values ​​are modified to preset generic placeholders or null values. The modified tag set is considered the second set of data, which contains the de-identified tag set.

[0034] Finally, based on this second data, the pixel data in the DICOM data stream is compressed to generate a processed DICOM file.

[0035] Specifically, after the second data is generated, a compression module is invoked. This module is configured to extract the pixel data portion from the original DICOM data stream. This pixel data is then processed using a general compression algorithm, such as a dictionary-based or transform-based compression algorithm. The compressed pixel data is then recombined with the desensitized tag set contained in the second data to form a new file conforming to the DICOM standard. This new file is the processed DICOM file.

[0036] In some possible embodiments, the step of parsing the entire set of raw tags in the DICOM data stream to generate first data containing all raw plaintext metadata includes: extracting all DICOM tags in the DICOM data stream; serializing all DICOM tags in plaintext JSON format; generating a separate metadata file and permanently storing all raw plaintext metadata.

[0037] Extracting all DICOM tags from the DICOM data stream refers to identifying and parsing all non-pixel data portions, i.e., various attribute tags, from the received DICOM data stream. These tags are typically encoded according to the format specified by the DICOM standard. Each tag contains a group number, element number, value representation (VR), value length, and the actual data value. This step is the first step in obtaining the raw metadata of the DICOM image, aiming to ensure that subsequent processing can be based on complete and accurate raw information. In practice, this can be achieved by traversing the byte sequence of the DICOM data stream, identifying and reading the identifier of each tag and its corresponding data content one by one according to the DICOM data dictionary and encoding rules.

[0038] Serializing all DICOM tags into plaintext JSON format means converting all extracted DICOM tags and their corresponding values ​​into a structured, easy-to-read, and machine-parseable text format, namely JSON (JavaScript Object Notation). The advantage of plaintext JSON format lies in its strong human readability and widespread support, facilitating data exchange between different systems. Through serialization, the complex DICOM tag structure is mapped into key-value pairs; for example, the group number and element number of a DICOM tag can serve as the key of a JSON object, and its value as the corresponding value. This allows the original metadata to retain its structured characteristics even after being separated from the DICOM file itself, making it easy for other applications or services to understand and process.

[0039] Generating a separate metadata file and permanently retaining all the original plaintext metadata refers to writing the serialized plaintext JSON data into a separate file, independent of the original DICOM file, after the DICOM tags are extracted and JSON serialized. This file typically uses the `metadata` extension to clearly identify its metadata file attributes. The purpose of permanent retention is to ensure that even if the original DICOM file undergoes de-identification, compression, or other processing that alters its content, its original, unmodified metadata can still be completely preserved for subsequent auditing, tracing, data analysis, or recovery needs. This independent storage method gives the original metadata lifecycle management capabilities independent of the DICOM image file, improving data security and availability. For example, this separate metadata file can be stored in the same storage path as the processed DICOM file and associated via filename, or stored in a dedicated metadata storage service.

[0040] In some possible embodiments, when the tags in the DICOM data stream are de-identified based on the first data to generate second data containing the de-identified tag set, the de-identified tag set is written into the file header of the DICOM data stream.

[0041] Specifically, after de-identifying the tags in the DICOM data stream, the resulting privacy-protected tag information is directly embedded into the standard header of the DICOM data stream. The DICOM header is a predefined area in the DICOM file structure used to store various metadata (such as patient name, examination date, image acquisition parameters, etc.). By writing the de-identified tag set into the header, it ensures that any system subsequently processing or viewing the DICOM file can only access the privacy-protected information, effectively preventing accidental leakage of sensitive data. This operation ensures that the de-identified DICOM file remains a valid file that conforms to the DICOM standard and can be parsed and used normally.

[0042] Building upon this foundation, this embodiment also proposes a dual metadata architecture that separates the original plaintext metadata from the anonymized tags. The core idea of ​​this architecture is to logically and physically separate the original, complete plaintext metadata, containing all unmodified sensitive information, from the anonymized tag set. This design aims to simultaneously meet the dual requirements of data privacy protection (through anonymized tags) and data integrity preservation (through the original plaintext metadata). The original plaintext metadata is irreplaceable for data auditing, traceability, and the recovery of original information in specific authorized scenarios. The anonymized tags are used to generate DICOM files for external provision or use in non-sensitive environments. This separation allows for the implementation of different security policies and access permissions for different types of metadata, thereby significantly improving overall data security and management flexibility.

[0043] Specifically, the original plaintext metadata is stored in a separate metadata file, while the anonymized tags are stored within the DICOM file. This means that the plaintext metadata containing the original sensitive information is stored in a separate `metadata` file. This file is typically managed separately from the anonymized DICOM file and can be protected with stricter access controls, encryption measures, or a more restricted storage environment to ensure its confidentiality. Meanwhile, the anonymized tag set is stored directly as part of the standard file header within the DICOM file. This clear separation of storage locations allows the anonymized DICOM file to exist independently and be used normally, providing necessary image information without relying on external files, while ensuring that its content complies with privacy protection requirements. This clear storage distinction further enhances the practicality and security of the dual metadata architecture.

[0044] In some possible embodiments, the step of desensitizing the tags in the DICOM data stream includes: reading a pre-configured list of sensitive tags; traversing the tags in the DICOM data stream and matching them with the list of sensitive tags; and performing a preset replacement operation on the successfully matched sensitive tags.

[0045] First, the system reads a pre-configured list of sensitive tags. This list is a predefined set of tags used to identify tags in the DICOM data stream that require de-identification. It can be stored as a structured file (e.g., XML, JSON, or CSV), in a database, or loaded into memory at system startup. Each entry in the list typically contains a unique identifier for the DICOM tag (such as a group number and element number, e.g., (0010, 0010) representing a patient's name), and the de-identification rule or replacement operation type corresponding to that tag can be further specified. Pre-configuration allows for flexible definition and management of the scope of sensitive information requiring de-identification.

[0046] Secondly, after receiving the DICOM data stream and parsing its tag set, the system iterates through the tags in the DICOM data stream and matches them against the sensitive tag list. This process involves comparing the identifier of the current tag with all entries in the pre-configured sensitive tag list. The matching algorithm can use exact matching, requiring the tag's group number and element number to be exactly the same; it can also support fuzzy matching based on regular expressions or wildcards to accommodate more complex de-identification needs. This step ensures that all tags in the DICOM data stream are systematically examined, thereby identifying potentially sensitive information.

[0047] Finally, once a DICOM tag is identified as a sensitive tag (i.e., it matches an entry in the sensitive TAG list), the system will perform pre-defined replacement operations on the matched sensitive tag. These operations aim to eliminate or obfuscate the original sensitive information while preserving the data structure and some usability as much as possible. Common replacement operations include, but are not limited to: clearing the tag value (e.g., setting it to a zero-length string), replacing it with a fixed placeholder (e.g., "[Desensitized]"), hashing the tag value to generate an irreversible hash value, randomly offsetting date or time information, or generalizing a specific value to a range value. These pre-defined operations ensure the consistency and automation of the desensitization process.

[0048] In some possible embodiments, a dynamic configuration interface is provided, through which tags in the sensitive TAG list can be added or deleted in a custom manner, and the sensitive TAG list can be updated and the changes taken effect.

[0049] Specifically, a dynamic configuration interface (MAC) is an interactive mechanism that allows users or system administrators to adjust system parameters and configurations while the system is running, without modifying code or restarting the service. This interface can be a graphical user interface (GUI), a command-line interface (CLI), or an application programming interface (API), such as a RESTful API or gRPC interface. Its core function is to provide a flexible entry point, making the management of the sensitive TAG list no longer static but real-time operable. Through the dynamic configuration interface, users can perform custom addition operations, i.e., inputting new DICOM tags (e.g., using their group number and element number, such as (0010, 0010) representing a patient's name) and specifying corresponding desensitization rules (e.g., replacing with a fixed string, clearing, hashing, or partial masking). The system will validate the input tag's format and add it to the sensitive TAG list. Simultaneously, users can also perform custom deletion operations, i.e., selecting one or more existing DICOM tags from the current sensitive TAG list and removing them from the list through the interface. These operations allow for fine-grained management and adjustment of the desensitization strategy according to actual needs. After adding or deleting items from the sensitive TAG list, the system can immediately or within a very short time apply these changes to the running DICOM data processing flow without interrupting service; that is, the sensitive TAG list is updated and takes effect. This can be achieved through various mechanisms, such as the system monitoring changes to the configuration file and automatically hot-loading it, or notifying the de-identification module to reload the latest sensitive TAG list via an internal messaging mechanism. Ensuring that updates take effect immediately is crucial for maintaining the continuity of data processing and the real-time nature of policies.

[0050] In some possible embodiments, the step of compressing the pixel data in the DICOM data stream based on the second data includes: separating the metadata segment and the pixel data segment in the DICOM data stream; preserving the plaintext format of the metadata segment without compression; and performing JPEG-LS lossless compression on the pixel data segment.

[0051] Specifically, when processing DICOM data streams, the first step is to perform structural analysis to separate the metadata segment from the pixel data segment. A DICOM file typically consists of a file header, file metadata, and a series of data elements, where each data element represents different information based on its tag and value. The metadata segment mainly contains data elements describing non-image data such as image features, patient information, and examination information, including patient name, examination date, equipment information, and image size. The pixel data segment specifically refers to the data elements storing the actual image data, typically (7FE0,0010)Pixel Data. The separation operation involves parsing the structure of the DICOM data stream to identify and logically or physically distinguish between these two parts of data. This can be achieved by traversing the DICOM data elements and determining whether each element belongs to metadata or pixel data based on a predefined DICOM data dictionary. In memory, pointers or data structures can be used to manage these two parts of data, or they can be copied to different buffers for independent processing.

[0052] After separating the metadata segment, this embodiment chooses to retain the plaintext format of the metadata segment without performing any compression. Although the metadata segment has a relatively small data volume, its content is crucial for image indexing, retrieval, display, and subsequent diagnostic processes. If the metadata is compressed, decompression must be performed every time this information is needed, which will introduce additional computational overhead and time latency, especially in scenarios where metadata needs to be frequently accessed for querying or previewing. Therefore, maintaining the plaintext format of the metadata segment ensures that it can be read and parsed quickly and directly by the system, greatly improving the accessibility and processing efficiency of the metadata, thereby accelerating the response speed of the image management system.

[0053] Meanwhile, for the pixel data segments that occupy the main part of the DICOM file, this embodiment performs JPEG-LS lossless compression processing. Pixel data segments typically contain a large amount of image data and are the main contributor to the size of the DICOM file. Using the JPEG-LS (Joint Photographic Experts Group - Lossless) lossless compression standard can significantly reduce the file size of pixel data segments without losing any original image information. Lossless compression is crucial for medical imaging because any change in pixel values ​​can affect the accuracy of diagnosis. JPEG-LS achieves a high lossless compression ratio through techniques such as prediction, context modeling, and entropy coding. In implementation, a standard JPEG-LS encoding library or algorithm module can be called, taking the original pixel data as input, encoding it to generate a compressed bitstream, and then replacing the pixel data segments in the original DICOM data stream with this compressed pixel data.

[0054] In some possible embodiments, the step of performing JPEG-LS lossless compression on the above-mentioned pixel data segment includes: reading the original pixel values ​​of the pixel data segment; performing lossless encoding on the original pixel values ​​using the JPEG-LS encoding algorithm; and replacing the original pixel data segment in the DICOM data stream with the compressed pixel data segment.

[0055] Specifically, when reading the raw pixel values ​​of the pixel data segment, the system parses the start position and length of the pixel data segment according to the DICOM standard, and loads it into memory according to the transfer syntax and pixel data type (such as unsigned integer or signed integer). This step aims to accurately extract the raw pixel information of the image from the DICOM data stream, which is the foundation for any image processing (including compression), ensuring that subsequent encoding operations can directly apply to the unprocessed raw image data, thereby guaranteeing lossless compression.

[0056] Subsequently, the original pixel values ​​are losslessly encoded using the JPEG-LS encoding algorithm. JPEG-LS is a highly efficient lossless image compression standard, particularly suitable for scenarios with extremely high image quality requirements, such as medical imaging. It achieves a high compression ratio while ensuring that pixel values ​​remain completely undistorted through predictive coding and context modeling techniques. Using the JPEG-LS encoding algorithm means that the system will call the corresponding encoding library or the logic implementing the algorithm to process the read original pixel values ​​pixel by pixel or line by line, generating a compressed bitstream. This lossless encoding method ensures the accuracy of medical image diagnosis and avoids any information loss introduced by compression.

[0057] Finally, the compressed pixel data segments replace the original pixel data segments in the DICOM data stream. This step aims to reintegrate the losslessly compressed pixel data back into the DICOM data stream to form a complete new file conforming to the DICOM standard. During the replacement process, it is necessary to update the metadata related to the pixel data in the DICOM file header. For example, the transfer syntax (UID) may need to be updated from the original uncompressed transfer syntax to the JPEG-LS lossless transfer syntax. The length information of the pixel data elements also needs to be updated.

[0058] In some possible embodiments, one scheme for optimizing the timing of DICOM medical image data transmission is to transmit the metadata file containing all original plaintext metadata immediately after the generation of the first data is completed; and to transmit the compressed and desensitized DICOM image file after the generation of the processed DICOM file is completed.

[0059] The phrase "immediately after generating the first data, transmit the metadata file containing all original plaintext metadata" means that after all the original tag sets parsed from the DICOM data stream are serialized in plaintext JSON format and stored in a separate metadata file, the transmission process of this metadata file is initiated immediately without waiting for subsequent de-identification and compression processing. This can be achieved through an asynchronous transmission mechanism. For example, after generating the metadata file, it can be placed in a separate transmission queue, and a dedicated transmission module can be responsible for sending it to the target receiving end. This transmission can use a standard file transfer protocol or be pushed through a message queue service. The purpose is to ensure that the metadata file containing the original plaintext metadata arrives at the receiving end as quickly as possible so that the receiving end can process and utilize the data as soon as possible.

[0060] "Transmitting the compressed and desensitized DICOM image file only after the generated DICOM file has undergone tag desensitization and pixel data compression" means that the transmission of the DICOM image file is initiated only after the DICOM file has been generated following tag desensitization and pixel data compression. This means that the transmission of this file occurs later than the transmission of the metadata file. The transmission of the DICOM image file is only initiated after the entire DICOM data stream parsing, desensitization, and compression processes are completed. This is typically done via the standard DICOM C-STORE SCU protocol, sending the processed DICOM file to a PACS system or other image archiving system. This transmission strategy ensures that the DICOM image file is transmitted only after its content is complete and meets security and storage requirements, while also coordinating with the priority transmission of metadata files to optimize overall data flow efficiency.

[0061] In some possible embodiments, the transmission of the metadata file containing all original plaintext metadata includes: first transmitting JSON plaintext metadata; the receiving end performing image retrieval and index construction in advance based on the JSON plaintext metadata; and subsequently receiving the compressed and desensitized DICOM image file and completing data association.

[0062] Specifically, the priority transmission of JSON plaintext metadata here aims to ensure that the metadata file containing all original plaintext metadata arrives at the receiving end as quickly as possible. JSON plaintext metadata refers to data serialized from the complete set of original tags in the DICOM data stream in an easily parsed and processed JSON format. Prioritized transmission can be achieved in several ways, such as initiating the transmission of the metadata file immediately before the data transmission of the DICOM image file begins; or, through a dedicated, high-priority network channel or message queue mechanism. This prioritization allows the receiving end to obtain critical metadata information as early as possible, buying time for subsequent operations.

[0063] Based on this, the receiving end performs image retrieval and index construction in advance using the JSON plaintext metadata. Upon receiving the prioritized JSON plaintext metadata, the receiving end can immediately parse and process it. This metadata contains key information such as patient name, ID, examination date, examination type, image modality, and sequence description. The receiving end can utilize this information to begin performing operations such as image retrieval, database index construction, and updating the worklist or user interface display even before receiving the complete DICOM image file. For example, the system can pre-create records in the database based on patient ID or examination date, or display a preliminary image list to the doctor's workstation, thereby significantly reducing user waiting time.

[0064] Subsequently, the system will receive the compressed and anonymized DICOM image files and complete the data association. After the receiving end completes the preprocessing based on JSON plaintext metadata, the system will continue to receive the previously compressed and anonymized DICOM image files. Once the DICOM image file transmission is complete, the receiving end will use its internal unique identifier (such as SOPInstance UID) to match and associate it with the previously received and indexed JSON plaintext metadata. In this way, the complete image data and the corresponding original metadata are accurately combined to form a complete DICOM object that can be viewed, diagnosed, and stored later.

[0065] In some possible embodiments, the specific method of real-time processing of the received DICOM data stream in the DICOM C-STORE SCP receiving link before the DICOM file is stored includes: intercepting the DICOM data stream in the C-STORE request; completing all parsing, desensitization, and compression processing in memory; and writing the processed DICOM file to the storage medium after processing is completed.

[0066] Specifically, when the DICOM C-STORE SCP receiving link receives a C-STORE request, it first intercepts the DICOM data stream in that request. The interception operation aims to ensure that the system has complete control and access to the data stream before it is written to any storage medium. This can be achieved by deploying a proxy module or interceptor at the DICOM service layer or network transport layer, which captures the data before it reaches the regular file write logic.

[0067] Subsequently, the intercepted DICOM data stream is loaded into memory, where all parsing, de-identification, and compression processes are completed. This means that all intermediate processing steps—including parsing the entire set of original tags in the DICOM data stream to generate first data containing all original plaintext metadata, de-identifying the tags in the DICOM data stream based on the first data to generate second data containing the de-identified tag set, and compressing the pixel data in the DICOM data stream based on the second data—are performed in the computer's random access memory (RAM). For example, the parsed tag information is directly stored in a data structure in memory, the de-identification operation directly modifies the tag data in memory, and the compression of pixel data is also completed in memory, with the compressed data stored in a memory buffer. This process avoids any disk I / O operations, thereby significantly improving processing speed and efficiency.

[0068] Ultimately, a one-time disk write operation is only performed after the DICOM data stream has undergone a complete parsing, desensitization, and compression process, and a final processed DICOM file data block has been generated in memory. This complete and compliant processed DICOM file data block is then written to the designated storage medium. This "process-before-write" strategy ensures that the data written to the storage medium is the final, complete, and compliant version.

[0069] It should be noted that the above description describes some embodiments of the present invention. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than that shown in the above embodiments and still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] Based on the same inventive concept, corresponding to any of the above embodiments, such as Figure 2 As shown in the embodiments, this application also discloses a device for desensitization and lossless compression processing of multimodal data acquisition in a medical imaging big data platform, comprising: The receiving and processing unit is used to process the received DICOM data stream in real time during the DICOM C-STORE SCP receiving link, before the DICOM file is stored on the ground. The parsing unit is used to parse the entire set of raw tags in the DICOM data stream and generate the first data containing all raw plaintext metadata. The desensitization processing unit is used to desensitize the tags in the DICOM data stream based on the first data, and generate second data containing the desensitized tag set; The compression processing unit is used to compress the pixel data in the DICOM data stream based on the second data to generate a processed DICOM file.

[0071] The apparatus described above is used to implement the desensitization and lossless compression processing method of the corresponding medical imaging big data platform in the multimodal data acquisition process in the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0072] Based on the same inventive concept, and corresponding to the methods of any of the above embodiments, this application also discloses an electronic device, such as... Figure 3 As shown, it includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the aforementioned desensitization and lossless compression processing method of the medical imaging big data platform during the multimodal data acquisition process.

[0073] Specifically, the device includes: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected within the device via the bus 1050.

[0074] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), GPU (Graphics Processing Unit), or one or more integrated circuits, to implement relevant programs and achieve the technical solutions provided in the embodiments of this specification.

[0075] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1020 and called by the processor 1010. The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, projectors, speakers, vibrators, indicator lights, etc.

[0076] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB (Universal Serial Bus), network cable, etc.) or wireless means (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).

[0077] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0078] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0079] Based on the same inventive concept, corresponding to any of the above embodiments, this application also discloses a computer-readable storage medium that stores computer instructions for enabling a computer to implement the desensitization and lossless compression processing method of the above-described medical imaging big data platform in the multimodal data acquisition process.

[0080] The computer-readable storage medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information accessible by a computing device. The computer instructions stored in the storage medium of the above embodiments are used to enable the computer to implement the desensitization and lossless compression processing method of the medical imaging big data platform in the multimodal data acquisition process as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0081] The above description is merely a preferred embodiment of the present invention and the technical principles employed. The present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the claims.

Claims

1. A method for de-sensitization and lossless compression of medical image big data platform in a multi-modal data acquisition process, characterized in that, include: In the DICOM C-STORE SCP receiving link, before the DICOM file is stored, the received DICOM data stream is processed in real time; Parse the complete set of raw tags in the DICOM data stream to generate first data containing all raw plaintext metadata; Based on the first data, the tags in the DICOM data stream are de-identified to generate second data containing a set of de-identified tags; Based on the second data, the pixel data in the DICOM data stream is compressed to generate a processed DICOM file.

2. The method for desensitization and lossless compression processing of multimodal data acquisition in the medical imaging big data platform according to claim 1, characterized in that, The process of parsing the entire set of raw tags in the DICOM data stream to generate first data containing all raw plaintext metadata includes: Extract all DICOM tags from the DICOM data stream; Serialize all the DICOM tags in plaintext JSON format; Generate a separate metadata file to permanently store all the original plaintext metadata. 3.The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 1, characterized in that, The step of desensitizing the tags in the DICOM data stream based on the first data to generate second data containing a set of desensitized tags includes: Write the desensitized tag set into the header of the DICOM data stream; A dual metadata architecture is formed, which separates and stores the original plaintext metadata from the de-identified tags; The original plaintext metadata is stored in a separate metadata file, while the de-identified tags are stored inside the DICOM file.

4. The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 1, characterized in that, The desensitization process for the tags in the DICOM data stream includes: Read the pre-configured list of sensitive tags; Traverse the tags in the DICOM data stream and match them with the list of sensitive tags; Perform a preset replacement operation on the successfully matched sensitive tags.

5. The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 4, characterized in that, Also includes: Provides a dynamic configuration interface; The dynamic configuration interface allows for the addition or deletion of tags in the sensitive TAG list. Update the sensitive TAG list and make it effective.

6. The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 1, wherein, The step of compressing the pixel data in the DICOM data stream based on the second data includes: Separate the metadata segment and pixel data segment in the DICOM data stream; The plaintext format of the metadata segment is retained without compression. Performing JPEG-LS lossless compression on the pixel data segment includes: Read the raw pixel values ​​of the pixel data segment; The original pixel values ​​are losslessly encoded using the JPEG-LS encoding algorithm. Replace the original pixel data segment in the DICOM data stream with the compressed pixel data segment.

7. The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 1, characterized in that, Also includes: After the first data is generated, the metadata file containing all the original plaintext metadata is immediately transmitted, including: Prioritize transmitting plaintext JSON metadata; The receiving end performs image retrieval and index construction in advance based on the JSON plaintext metadata; The compressed and desensitized DICOM image file is then received and data association is completed. After the DICOM file is generated, the compressed and desensitized DICOM image file is then transmitted. 8.The method of desensitization and lossless compression of medical image big data platform in multi-modal data acquisition process according to claim 1, wherein, The real-time processing of the received DICOM data stream in the DICOM C-STORE SCP receiving link, before the DICOM file is stored, includes: Intercept the DICOM data stream in the C-STORE request; Complete all parsing, desensitization, and compression processes in memory; After processing is complete, the processed DICOM file is written to the storage medium. 9.A device for desensitization and lossless compression processing of medical image big data platform in a multi-modal data acquisition process, characterized in that, include: The receiving and processing unit is used to process the received DICOM data stream in real time during the DICOM C-STORE SCP receiving link, before the DICOM file is stored on the ground. The parsing unit is used to parse the entire set of original tags in the DICOM data stream and generate first data containing all original plaintext metadata. The desensitization processing unit is used to desensitize the tags in the DICOM data stream based on the first data, and generate second data containing a set of desensitized tags; The compression processing unit is used to compress the pixel data in the DICOM data stream based on the second data to generate a processed DICOM file.

10. An electronic device, comprising: include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to cause the at least one processor to perform the desensitization and lossless compression processing method of the medical imaging big data platform according to any one of claims 1 to 8 in the multimodal data acquisition process.

Images