DICOM image downloading and analyzing method based on adaptive demand

Abstract

The present application relates to the technical field of digital image, especially relates to a DICOM image downloading and analyzing method based on adaptive demand, a downloading service scheduling center is established between a downloading client and a downloading node; the downloading service scheduling center provides a discovery service for discovering storage nodes in the downloading node; the downloading client needs to register to the downloading service scheduling center when starting, and is registered as a downloading service sub-node; the downloading service scheduling center responds to the downloading request of the downloading service sub-node through dynamic scheduling. The present application dynamically adjusts the downloading and analyzing strategy according to user operation, and comprehensively considers the balance of performance and memory occupation in the process, greatly improves the DICOM image downloading and analyzing efficiency.

Description

Technical Field

[0001] This invention relates to the field of digital imaging technology, and in particular to a method for downloading and parsing DICOM images based on adaptive requirements. Background Technology

[0002] Digital imaging refers to the original, lossless compressed DICOM (Digital Imaging and Communications in Medicine) format images generated during X-ray computed tomography (CT) scans, magnetic resonance imaging (MRI) scans, and X-ray examinations. These images are stored on storage servers (SAN, NAS, etc.) and can be downloaded and viewed directly as many times as possible through a PACS (picture archiving and communication system) client.

[0003] In recent years, with the improvement of examination equipment technology and the need for epidemic diagnosis, thin-slice CT scanning technology has entered a stage of large-scale application. The number of images in a single CT examination can reach 1,000 to 10,000 (a single examination contains several sequences; a sequence contains several images). The dramatic increase in image data volume has brought great challenges to the back-end services of PACS systems, especially creating a very serious performance bottleneck for image retrieval services. On the other hand, traditional PACS system clients download images sequentially according to the examinations specified by the user. During the download process, sequential images cannot be manipulated, or examinations or sequences of interest must be queued and cannot be downloaded first. This results in excessively long waiting times for radiologists, significantly reducing their image interpretation efficiency and thus leading to a decrease in overall work efficiency. Summary of the Invention

[0004] This invention provides an adaptive DICOM image download and parsing method for PACS systems. It completely decouples the download and parsing processes through a planning algorithm and dynamically adjusts the download priority according to user operation requirements to complete the DICOM image download.

[0005] To achieve the objectives of this invention, the technical solution adopted is: a DICOM image download and parsing method based on adaptive requirements, comprising:

[0006] Establish a download service scheduling center between the download client and the download node;

[0007] The download service dispatch center provides a discovery service to discover storage nodes among the download nodes;

[0008] When the download client starts, it needs to register with the download service scheduling center and register as a sub-node of the download service;

[0009] The download service scheduling center responds to download requests from download service sub-nodes in the download client through dynamic scheduling.

[0010] As an optimized solution of the present invention, the downloading step includes:

[0011] 1) The download client creates a download thread pool for the download sequence, and the download thread pool requests the download node where the download sequence is located from the download service scheduling center;

[0012] 2) The download service scheduling center decomposes the download sequence to be downloaded. Using the ID of the decomposed download sequence as the query condition, it queries its own sequence routing table. If the result is not found, the download node is stored in the sequence dynamic routing table. If the result is found, the download node that is currently downloading and has the fewest downloads is returned as the sequence download node.

[0013] 3) The download client starts image downloading using the sequence download node as the download path.

[0014] As an optimization of the present invention, the number of threads in the download thread pool is determined by the number of download sequences, each download sequence corresponds to one thread, and each image to be downloaded corresponds to an independent task in the thread.

[0015] As an optimization of the present invention, each download node provides FTP, HTTP and DICMC-Store download methods, and each download node has its own priority configuration for the download method.

[0016] As an optimized solution of the present invention, after a sequence of downloads is completed, the download client updates the ID of the downloaded sequence, the download client tag, and the download completion tag to the download service scheduling center.

[0017] As an optimized solution of the present invention, the analysis method includes the following steps:

[0018] A. The film viewing client establishes a sequence parsing thread pool for parsing tasks;

[0019] B. After downloading, submit the data to the sequence parsing thread pool to perform the parsing task. When extracting the image description information data in the file, extract it layer by layer as needed, divide the information into four levels: patient, examination, sequence, and image, and organize them into each level respectively.

[0020] C. When a user is browsing a sequence of images, pause the parsing of other sequences, prioritize parsing the sequence currently being viewed, and load the current image pixel data to notify the UI to draw;

[0021] D. When rendering images, considering that CT and MR images are small in size but have a large number of images, while CR, DR, and MG images are large in size but have a small number of images, it is necessary to load image pixel data by comprehensively considering performance and memory usage.

[0022] E. After the current sequence is parsed, resume parsing the remaining sequences until all sequence images are parsed.

[0023] As an optimized solution of the present invention, the strategy for loading image pixel data is as follows:

[0024] When CT / MR images need to be displayed, the image pixel data is loaded from the file and drawn, without caching the pixel data;

[0025] DR / CR / MG image files are parsed and the pixel data is loaded immediately and cached in memory.

[0026] The present invention has positive effects: It provides a new method for downloading and parsing DICOM images, which dynamically adjusts the downloading and parsing strategy according to user operations, and comprehensively considers the balance between performance and memory usage in the process, which greatly improves the efficiency of DICOM image downloading and parsing. Attached Figure Description

[0027] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0028] Figure 1 This is an overall framework diagram of the present invention;

[0029] Figure 2 This is a schematic diagram of the download process of the present invention;

[0030] Figure 3 This is a schematic diagram of the analytical process of the present invention. Detailed Implementation

[0031] like Figure 1 As shown, this invention discloses a DICOM image download and parsing method based on adaptive requirements, comprising:

[0032] Establish a download service scheduling center between the download client and the download node;

[0033] The download service dispatch center provides a discovery service to discover storage nodes among the download nodes;

[0034] When the download client starts, it needs to register with the download service scheduling center and register as a sub-node of the download service;

[0035] The download service scheduling center responds to download requests from download service sub-nodes in the download client through dynamic scheduling.

[0036] like Figure 2 As shown, the download steps include:

[0037] 1) The download client creates a download thread pool for the download sequence, and the download thread pool requests the download node where the download sequence is located from the download service scheduling center;

[0038] 2) The download service scheduling center decomposes the download sequence to be downloaded. Using the ID of the decomposed download sequence as the query condition, it searches its own sequence routing table. If the result is not found, the download node is stored in the sequence dynamic routing table. If the result is found, the download node that is currently downloading and has the fewest downloads is returned as the sequence download node.

[0039] 3) The download client starts image downloading using the sequence download node as the download path.

[0040] An image is the smallest unit of download in a PACS system. Creating a download link for each image would lead to wasted resources and performance bottlenecks. A sequence is a collection of images, typically consisting of 1 to 10,000 images. The basic unit of advanced post-processing in CT and MR is a sequence. A single examination contains multiple sequences.

[0041] When the download service scheduling center starts up, it initializes the pre-configured storage nodes from the database configuration table and encapsulates different storage nodes so that they can be downloaded using the same download protocol.

[0042] When a radiologist opens a patient's examination image in the PACS system (download client), the PACS system creates a dedicated download thread pool for that examination. The first task of the thread pool is to request the download nodes containing the multiple sequences to be downloaded from the download service scheduling center. The download service scheduling center queries its own sequence routing table using the sequence ID as the query condition. If no data is found, it queries the database and stores the results in the dynamic sequence routing table of the download nodes before the database returns. If data is found, it selects and returns the node with the fewest currently downloading sequences. The dynamic sequence routing tables of the nodes that have already downloaded sequences update the download service scheduling center's sequence routing table. The download service scheduling center responds to download requests from download service child nodes through dynamic scheduling.

[0043] Once a sequence of downloads by the client is complete, the client updates the downloaded sequence ID, client tag, and download completion tag to the download service scheduling center, and the download task scheduling center reduces the download count for this node.

[0044] The download client can also set up a download management server:

[0045] 1) Establish a dynamic node management center for downloading between the download client and the physical storage node.

[0046] 2) Record the downloaded checks, sequence markers, client markers, and sequence download status for each download client.

[0047] 3) Based on the downloaded sequences and node location information, a distributed online downloadable node network is formed in real time.

[0048] 4) When the download client requests an inspection download, it performs a distributed search for downloadable nodes across all sequences under that inspection and returns the shortest download path to the download client. The download client performs the following download steps for each image inspection:

[0049] (1) Create a thread pool to perform a check on the total image download task.

[0050] When creating the thread pool, a thread-local object is created for each thread to simultaneously download the client connection. Each thread executes its image download task independently, without shared resource locking protection, thus improving download speed. Each check creates its own thread pool to execute the image loading tasks within that check. When a user opens multiple checks, later-opened checks do not need to wait for earlier checks to complete; each check executes in parallel. The number of threads in each check's thread pool is determined by the sequence number, with one thread corresponding to each sequence, and each image to be downloaded corresponding to an independent task (Job) within that thread.

[0051] (2) Each download node provides multiple download methods such as FTP, HTTP, and DICOM C-Store to ensure smooth and reliable task execution. Each node has its own priority configuration for the download method. When using the FTP download method, the download nodes are randomly assigned to achieve software load balancing. When a node becomes unavailable, the download can be dynamically switched to another node. When an FTP download node becomes unavailable, the download can automatically switch to the next priority download method, maximizing the reliability of the download task.

[0052] Image parsing / display process for downloading the client

[0053] The process of transferring an image from a remote storage system to a client interface involves not only file downloading but also file parsing, information extraction, and image rendering. When all these steps are performed sequentially in the same thread, the system's concurrency capabilities cannot be effectively utilized, and the time spent on file parsing and information extraction can impact subsequent image downloads. Therefore, the image downloading and parsing processes are separated and executed in different threads. A diagram illustrating the parsing process is shown below. Figure 3 As shown, the parsing method includes the following steps:

[0054] A. The film viewing client establishes a sequence parsing thread pool for parsing tasks;

[0055] B. After downloading, submit the data to the sequence parsing thread pool to perform the parsing task. When extracting the image description information data in the file, extract it layer by layer as needed, divide the information into four levels: patient, examination, sequence, and image, and organize them into each level respectively.

[0056] C. When a user is browsing a sequence of images, pause the parsing of other sequences, prioritize parsing the sequence currently being viewed, and load the current image pixel data to notify the UI to draw;

[0057] D. When rendering images, considering that CT and MR images are small in size but have a large number of images, while CR, DR, and MG images are large in size but have a small number of images, it is necessary to load image pixel data by comprehensively considering performance and memory usage.

[0058] E. After the current sequence is parsed, resume parsing the remaining sequences until all sequence images are parsed.

[0059] The strategy for loading image pixel data is as follows:

[0060] When CT / MR images need to be displayed, the image pixel data is loaded from the file and drawn, without caching the pixel data;

[0061] DR / CR / MG image files are parsed and the pixel data is loaded immediately and cached in memory.

[0062] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An adaptive demand based DICOM image download and parsing method, characterized in that: include: Establish a download service scheduling center between the download client and the download node; The download service dispatch center provides a discovery service to discover storage nodes among the download nodes; When the download client starts, it needs to register with the download service scheduling center and register as a sub-node of the download service; The download service scheduling center responds to download requests from download service sub-nodes in download clients through dynamic scheduling; The download steps include: 1) The download client creates a download thread pool for the download sequence, and the download thread pool requests the download node where the download sequence is located from the download service scheduling center; 2) The download service scheduling center decomposes the download sequence to be downloaded. Using the ID of the decomposed download sequence as the query condition, it searches its own sequence routing table. If the result is not found, the download node is stored in the sequence dynamic routing table. If the result is found, the download node that is currently downloading and has the fewest downloads is returned as the sequence download node. 3) The download client starts image downloading using the sequence download node as the download path; The parsing method includes the following steps: A. The film viewing client establishes a sequence parsing thread pool for parsing tasks; B. After downloading, submit the data to the sequence parsing thread pool to perform the parsing task. When extracting the image description information data in the file, extract it layer by layer as needed. Divide the information into four levels: patient, examination, sequence, and image, and organize them into each level respectively. C. When a user is browsing a sequence of images, pause the parsing of other sequences, prioritize parsing the sequence currently being viewed, and load the current image pixel data to notify the UI to draw; D. Image rendering requires a comprehensive consideration of performance and memory usage when loading image pixel data; E. After the current sequence is parsed, resume parsing the remaining sequences until all sequence images are parsed.

2. The DICOM image download and parsing method based on adaptive demand according to claim 1, characterized in that: The number of threads in the download thread pool is determined by the number of download sequences. Each download sequence corresponds to one thread, and each image to be downloaded corresponds to an independent task in the thread.

3. The DICOM image download and parsing method based on adaptive demand according to claim 2, characterized in that: Each download node offers FTP, HTTP, and DICOM C-Store download methods, and each download node has its own priority configuration for the download method.

4. The DICOM image download and parsing method based on adaptive demand according to claim 3, characterized in that: After a sequence of downloads is completed, the download client updates the ID of the downloaded sequence, the download client tag, and the download completion tag to the download service scheduling center.

5. The DICOM image downloading and parsing method based on adaptive requirements according to claim 4, characterized in that: The strategy for loading image pixel data is as follows: When CT / MR images need to be displayed, the image pixel data is loaded from the file and drawn, without caching the pixel data; DR / CR / MG image files are parsed and the pixel data is loaded immediately and cached in memory.

Images