WO2021169122A1 - 图像标注管理方法、装置、计算机系统及可读存储介质 - Google Patents

图像标注管理方法、装置、计算机系统及可读存储介质 Download PDF

Info

Publication number
WO2021169122A1
WO2021169122A1 PCT/CN2020/099291 CN2020099291W WO2021169122A1 WO 2021169122 A1 WO2021169122 A1 WO 2021169122A1 CN 2020099291 W CN2020099291 W CN 2020099291W WO 2021169122 A1 WO2021169122 A1 WO 2021169122A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
binary
user terminal
information
closed curve
Prior art date
Application number
PCT/CN2020/099291
Other languages
English (en)
French (fr)
Inventor
刘俊
刘晓洁
邢洁
陈超
黄凌云
刘玉宇
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2021169122A1 publication Critical patent/WO2021169122A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G06F16/538—Presentation of query results
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10132—Ultrasound image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30096—Tumor; Lesion

Definitions

  • This application relates to the field of computer technology, which relates to computer vision technology of artificial intelligence, and in particular to an image labeling management method, device, computer system, and readable storage medium.
  • the purpose of this application is to provide an image labeling management method, device, computer system, and readable storage medium, which are used to solve the existing technologies of low labeling efficiency, long time, uneven labeling quality, and lack of effectiveness for ultrasound images. It is not convenient for doctors in other departments to quickly and accurately understand the focus area and focus information in the image.
  • this application provides an image labeling management method, which includes the following steps:
  • Obtain the display request sent by the client obtain the file information matching the display request from the query tree, retrieve the target image corresponding to the file information, and send the target image and its file information to the Client display;
  • the selected area information is the range pulled or delineated on the target image by the user terminal;
  • the annotation label information sent by the user terminal and associate the annotation label information with the closed curve, convert the binary image into an annotation image, and store the annotation image in the database; where
  • the area delineated by the closed curve is the lesion area in the target image, and the label information is feature data describing the characteristics of the lesion area.
  • this application also provides an image labeling management device, including:
  • the creation module is used to create a database for storing at least one image and having a query tree, the query tree is associated with the file information of the image; wherein the image is a medical image, and the file information describes the attributes of the image The data;
  • the retrieval module obtains the display request sent by the client, obtains the file information matching the display request from the query tree, retrieves the target image corresponding to the file information, and compares the target image and its file The information is sent to the client for display;
  • the binary module is used to obtain the selected area information selected by the user terminal on the target image, and send a binary dialog box to the user terminal to obtain binary data, and perform binary data on the selected area information according to the binary data.
  • the binary image is obtained by quantization processing and sent to the user terminal for display; wherein, the selected area information is the range pulled or delineated on the target image by the user terminal;
  • the growth module is used to obtain the seed pixels and similarity thresholds selected by the user terminal on the binary image, perform region growth processing on the binary image to obtain a closed curve, and send the closed curve to the user terminal ;
  • the annotation module is used to obtain the annotation label information sent by the user terminal, and associate the annotation label information with the closed curve, convert the binary image into an annotated image, and store the annotated image in the A database; wherein the area delineated by the closed curve is a lesion area in the target image, and the label information is feature data describing the characteristics of the lesion area.
  • the present application also provides a computer system, which includes a plurality of computer devices, each computer device includes a memory, a processor, and a computer program stored in the memory and running on the processor.
  • each computer device includes a memory, a processor, and a computer program stored in the memory and running on the processor.
  • the processor of the device executes the computer program, the steps of the above-mentioned image labeling management method are jointly implemented.
  • the present application also provides a computer-readable storage medium, which includes multiple storage media, each of which stores a computer program, and when the computer program stored in the multiple storage media is executed by a processor Jointly realize the steps of the above-mentioned image labeling management method.
  • the image tagging management method, device, computer system, and readable storage medium create a database for storing images and having a query tree that is associated with the file information of the image to effectively perform the image Management, which eliminates the confusion of the current management system for image management; retrieves the target image and its file information from the database according to the display request through the query tree, and displays the target image and its file information through the user terminal; Binarize the selected area with binary data to obtain the user to reduce the interference of irrelevant image information to the doctor; perform area growth processing on the binary image through seed pixels and similarity thresholds to obtain a closed curve, the closed curve
  • the delineated area is the area of the lesion that the doctor needs to mark, which realizes the purpose of automatically and accurately drawing the closed curve, improves the drawing efficiency, shortens the drawing time, and at the same time guarantees the drawing and labeling quality of the closed curve;
  • the area delineated by the closed curve is colored to mark the area, so that doctors in other departments can quickly learn the range of the lesion area and its pathological characteristics
  • FIG. 1 is a flowchart of Embodiment 1 of the image labeling management method of this application;
  • Fig. 2 schematically shows an environmental application diagram of the image labeling management method according to the first embodiment of the present application
  • step S3 is a flowchart of sending a binary dialog box to the user terminal to obtain binary data in step S3 of the first embodiment of the image labeling management method of this application;
  • step S3 of the first embodiment of the image labeling management method of this application is a flowchart of performing binarization processing on the selected region information according to the binary data in step S3 of the first embodiment of the image labeling management method of this application to obtain a binary image;
  • FIG. 5 is a flowchart of a method for obtaining seed pixels and similarity thresholds in step S4 of the first embodiment of the image tagging management method of this application;
  • step S4 is a flowchart of performing region growth processing on a binary image to obtain a closed curve in step S4 of the first embodiment of the image labeling management method of this application;
  • Fig. 7 is a flowchart of associating the label information with the closed curve in step S5 of the first embodiment of the image labeling management method of this application;
  • step S5 is a flowchart of storing the annotated image in the database in step S5 of the first embodiment of the image annotation management method of this application;
  • FIG. 9 is a schematic diagram of the program modules of the second embodiment of the image labeling management device of this application.
  • FIG. 10 is a schematic diagram of the hardware structure of the computer equipment in the third embodiment of the computer system of this application.
  • Image labeling management device 2. Server 3. Network 4. Client
  • the image labeling management method, device, computer system, and readable storage medium provided in this application are applicable to the computer field, and provide an image labeling management method based on a creation module, a retrieval module, a binary module, a growth module, and an annotation module .
  • This application creates a database for storing images and has a query tree, the query tree is associated with the file information of the image; the query tree is used to retrieve the target image and its file information from the database according to the display request, and The target image and its file information are displayed through the user terminal; the selected area is binarized through the binary data; the binary image is subjected to region growth processing through the seed pixels and similar thresholds to obtain a closed curve.
  • the area delineated by the curve is the lesion area that the doctor needs to mark; the area delineated by the closed curve is color-marked by the marking label.
  • An image labeling management method of this embodiment includes the following steps:
  • S1 Create a database for storing at least one image and having a query tree, the query tree is associated with file information of the image; wherein the image is a medical image, and the file information is data describing image attributes;
  • S2 Obtain the display request sent by the client, obtain the file information matching the display request from the query tree, retrieve the target image corresponding to the file information, and send the target image and its file information Said user terminal display;
  • S3 Obtain the information of the selected area selected on the target image by the user end, and send a binary dialog box to the user end to obtain the binary data, and perform binarization processing on the selected area information according to the binary data. Binary image and send it to the user terminal for display; wherein, the selected area information is the range pulled or delineated on the target image by the user terminal;
  • S4 Obtain the seed pixel and similarity threshold selected by the user terminal on the binary image, perform region growth processing on the binary image to obtain a closed curve, and send the closed curve to the user terminal;
  • S5 Obtain annotated label information sent by the client, and associate the annotated label information with the closed curve, convert the binary image into an annotated image, and store the annotated image in the database; wherein The area encircled by the closed curve is the lesion area in the target image, and the label information is feature data describing the characteristics of the lesion area.
  • Fig. 2 schematically shows an environmental application diagram of the image labeling management method according to the first embodiment of the present application.
  • the image labeling management method runs in the server 2, and a database for storing images and having a query tree is created in the server 2; the server 2 is connected to multiple clients 4 through the network 3.
  • the query tree includes first-level nodes, second-level nodes, and third-level nodes; the file information includes first-level information and second-level information.
  • the first-level information is information describing the image owner data
  • the second-level Information is information describing the basic situation of the image based on primary information; preferably, the primary information is patient information, and the secondary information is the examination time based on the patient information; for example, the patient information may include but Not limited to: name, age, ID number; wherein, the inspection time is the time when the patient took the image, which can be displayed in the form of year/month/day.
  • the first-level information is stored in a first-level node
  • the second-level information is stored in a second-level node corresponding to the first-level information
  • the image is stored in a second-level node corresponding to the second-level information.
  • the third-level node the association between the query tree and the file information is realized;
  • the query tree is a file tree list created based on QTreeWidget, which displays the file information according to three-level nodes; among them, QTreeWidget is a tree-shaped component that displays various items in the form of a tree. Each item is represented by QTreeWidgetItem.
  • the dialog box is made based on QDialog.
  • the QDialog class is the base class of the dialog window.
  • the dialog window is the top-level window mainly used for short-term tasks and brief communication with users.
  • QDialog can be modal or non- Mode
  • QDialog supports extensibility and can provide return values.
  • the examination time stored under the second-level node is arranged in ascending or descending order of the examination time, which is convenient for the doctor to select and mark.
  • the display request includes a first request for describing the information of the image owner and a second request for describing the basic situation of the image; the server 2 obtains information from the first-level node of the query tree according to the display request. Request matching first-level information, obtain second-level information matching the second request from second-level nodes under said first-level information, and extract images stored in third-level nodes under said second-level information, And set it as the target image; send the target image and the file information of the target image to the client for display.
  • the third-level node also stores diagnostic information
  • the diagnostic information is a diagnosis conclusion made by a professional doctor on the image lesion, which is data information after structured processing, which can be compared with the target image
  • the user terminal display is sent together so that it can be output to the user terminal and displayed in the patient information column; wherein the file information and diagnostic information are displayed on the user terminal through the patient information column, so that the doctor can refer to the pathology table information Labeling can provide the labeling doctor with more information besides the image, which is convenient for the doctor to make more accurate labeling.
  • initialize the patient information bar to restore it to a blank state, so as to display the file information and diagnostic information.
  • the user terminal 4 can obtain the selected area information by dragging the mouse on the target image.
  • the server 2 recognizes the selected area information and sends a binary dialog box to the user terminal 4.
  • the user can use the user terminal 4 in the binary dialog Enter the binary data in the box to adjust the parameters of the binary processing operation of the server 2; the server 2 performs binary processing on the selected area information according to the binary data to obtain the binary image and send it to the client 4 show. Since the range of the binarization processing is obtained through the selected region information, it is only necessary to perform the binarization processing on the required range, without performing the overall binarization processing on the target image.
  • the user can click on a certain pixel in the binary image through the user terminal 4, or pull a rectangular frame with at least one pixel to obtain the seed pixel; the server 2 sends a similar dialog box to the user terminal 4 according to the seed pixel, and the user can The user terminal 4 inputs the similarity threshold in the similarity dialog box to adjust the parameters of the region growth processing operation of the server 2; the server 2 performs region growth processing on the binary image according to the similarity threshold and the seed pixels to obtain similar regions.
  • the boundary of the similar area draws a closed curve, and sends it to the user terminal 4 for display.
  • the server 2 sends a label dialog box to the client 4 according to the label request triggered by the client 4.
  • the client 4 selects the pathology label in the label dialog box and sends it as the label information to the server 2, and the server 2 according to the label information
  • the label color of is assigned color to the area delineated by the closed curve. At this time, the binary image will be converted into an annotated image, and the annotated image will be stored in the database.
  • Network 3 can include various network devices, such as routers, switches, multiplexers, hubs, modems, bridges, repeaters, firewalls, proxy devices and/ Or wait.
  • the network 3 may include physical links, such as coaxial cable links, twisted pair cable links, optical fiber links, combinations thereof, and/or the like.
  • the network 3 may include wireless links, such as cellular links, satellite links, Wi-Fi links, and/or the like.
  • the server 2 may be composed of a single or multiple computer devices (eg, servers).
  • the single or multiple computing devices may include virtualized computing instances.
  • Virtualized computing instances may include virtual machines, such as computer system simulations, operating systems, servers, and so on.
  • the computing device may load the virtual machine based on a virtual image and/or other data defining specific software (e.g., operating system, dedicated application, server) for simulation.
  • specific software e.g., operating system, dedicated application, server
  • a hypervisor can be implemented to manage the use of different virtual machines on the same computing device.
  • obtaining the file information matching the display request from the query tree in S2, and invoking the target image corresponding to the file information includes:
  • S201 Extract the first request in the display request, and obtain the first-level information matching the first request from the first-level nodes of the query tree, and obtain the second-level node corresponding to the first-level information .
  • the first-level information matching "Zhang San” is obtained from the first node of the query tree, such as: Zhang San or the name Zhang San, and the query tree is obtained The second-level node under "Zhang San or name Zhang San” in the middle.
  • S202 Extract a second request in the display request, and obtain second-level information matching the second request from the second-level node, obtain a third-level node corresponding to the second-level information, and Set the primary information and secondary information as file information.
  • the secondary information matching "2018-01-20” is obtained from the second node of the query tree, such as: 2018-01-20 or 2018 On January 20, 2018, and obtain the third-level node under "2018-01-20 or 2018-01-20” in the query tree, and use Zhang San and 2018-01-20 as the file information.
  • DICOM Digital Imaging and Communications in Medicine
  • ISO 12052 International Standard for medical images and related information
  • Pydicom is a pure Python software package for processing DICOM files, which can be passed Very easy "Pythonic" way to extract and modify DICOM data, the modified data will also generate new DICOM files.
  • the selected area information in S3 is that the user terminal uses a cursor to delineate a selected area on the target image, and the generated information used to describe all pixels in the area includes at least each The color information, gray value information and coordinate information of the pixel.
  • the sending of a binary dialog box to the user terminal in S3 to obtain binary data includes:
  • S301 Send a binary dialog box to the user terminal; wherein the binary dialog box has an upper limit input box and a lower limit input box;
  • S302 Receive binary data generated by the user terminal by filling in a binary dialog box; wherein, the binary data is the upper processing limit threshold and the lower processing limit threshold input by the user terminal in the upper limit input box and the lower limit input box;
  • the upper limit input box and the lower limit input box on the binary dialog box can be set by a SliderBar (slider control), and the user terminal can obtain the processing upper threshold, the processing lower threshold, and similar thresholds by adjusting the SliderBar.
  • SliderBar sliding control
  • performing binarization processing on the selected region information according to the binary data in S3 to obtain a binary image includes:
  • S311 Extract the gray values of all pixels in the selected area information
  • S312 Set a pixel with a gray value greater than the processing upper threshold as an over-dark pixel, and increase the gray value of the over-dark pixel, for example, adjust the gray value of the all-black pixel to 255;
  • S313 Set a pixel with a gray value less than the processing lower limit threshold as an over-bright pixel, and lower the gray value of the over-bright pixel, for example, adjust the gray value of the all-white pixel to 0.
  • the noise reduction of the selected area information is realized to reduce the interference of irrelevant image information to the doctor, so that the doctor can accurately identify the useful information in the binary image.
  • the binary image is superimposed on the original image in the form of overlay (a virtualization technology mode that superimposes on the basis of the original image), which is not only convenient for doctors to compare and observe, but also does not require modification to the original image. , To avoid the unrecoverable situation of the original picture due to modification.
  • the seed pixels and similar thresholds in S4 can be obtained in the following manner:
  • S401 Receive a pixel clicked or circled on the binary image by the user terminal, and set the pixel as a seed pixel;
  • create a seed stack for storing pixels receive the coordinate data generated by the user terminal by clicking or circle on the binary information image, set the pixel corresponding to the coordinate data as the seed pixel, and store it To the seed stack.
  • S402 Send a similar dialog box to the user terminal according to the seed pixels, and receive the similarity threshold input by the user terminal in the similar dialog box.
  • the similar input box on the similar dialog box can be set by means of SliderBar (slider control), and the user terminal can obtain the similarity threshold by adjusting the SliderBar.
  • performing region growth processing on the binary image in S4 to obtain a closed curve includes:
  • S411 Extract the gray value of the seed pixel, and obtain a similarity interval according to the gray value and a similarity threshold;
  • extract the gray value of the seed pixel and set the gray value as a calculated value wherein if there is one seed pixel, extract the gray value of the seed pixel and set it as the calculated value. If the number of seed pixels is two or more, then extract the gray values of various sub-pixels, add each gray value and divide by the number of seed pixels to obtain the average gray value. The average gray value is set as the calculated value;
  • the calculated value and the similarity threshold are added to obtain the upper limit of similarity, the calculated value and the similarity threshold are subtracted to obtain the lower limit of similarity, and the similarity interval with the upper limit of similarity and the lower limit of similarity as the upper limit and the lower limit is compiled.
  • S412 Set pixels with gray values in the similar interval in the binary image as similar pixels, and store the seed pixels and similar pixels in a preset seed stack;
  • the gray values of pixels in the binary image are extracted, pixels whose gray values belong to a similar interval are set as similar pixels, and they are stored in the seed stack.
  • S413 Extract pixels located at the boundary in the seed stack and set them as boundary pixels, and draw a closed curve on the binary image along the boundary pixels.
  • drawing a closed curve on the binary image along the boundary pixels in the step S413 further includes:
  • the label information can be obtained in the following manner:
  • S501 Pre-create a selection database for storing label dialogs with pathology labels.
  • a selection database is created in advance to store the label dialog box, and pathology labels are formulated in the label dialog box according to user needs.
  • the pathology label may include: calcification, mass, structural distortion, asymmetry, etc., used to describe breasts The pathology name of the lesion; by setting a selection box or a check box on the pathology label, it is loaded in the label dialog box, so that the user can select the desired pathology label by clicking the selection box or the check box.
  • the label dialog box includes a thyroid dialog box and a breast dialog box;
  • the thyroid dialog box includes: thyroid disease type label, thyroid internal composition label, internal strong echo label, echo type label, shape label, border label, edge label, All icons are marked with label information;
  • the label dialog box is a breast dialog box including: breast disease type label, BI-RADS label, benign and malignant label, echo mode label, edge label, orientation label, and shape label; then the user terminal is received The sub-tab of the pathological label selected in the breast dialog box.
  • S502 Extract a label dialog box from the selection database according to the closed curve, and output it to the user terminal to make it pop up on the user terminal.
  • the label dialog box is a pre-defined template for disease selection based on QDialog.
  • the QDialog class is the base class of the dialog window, and the dialog window is mainly used for short-term tasks and brief communication with users.
  • the top-level window, QDialog can be modal or non-modal, QDialog supports scalability and can provide return values; through the use of tabbed dialog box to avoid the doctor’s description of the lesion varies from person to person, thus avoiding other doctors’ reasons Misunderstanding of the doctor's label, resulting in misdiagnosis.
  • S503 Receive the pathology label selected by the user terminal in the label dialog box to obtain label label information.
  • the user selects the desired sub-label of the pathology label by clicking the selection box or the check box to obtain label information, so that the doctor can define or mark the area delineated by the closed curve;
  • the closed curve outputs a category dialog box with disease type options to the user terminal and makes it pop up on the user terminal.
  • the disease type options include thyroid options and breast options; if the user terminal selects the thyroid option, output thyroid to the user terminal Dialog box; if the user endpoint selects the breast option, the breast dialog box is output to the user terminal; the user terminal can select the sub-tabs that require pathology labels in the thyroid dialog box or the breast dialog box to obtain descriptions
  • the label information of the lesion feature in the area enclosed by the closed curve are examples of the lesion feature in the area enclosed by the closed curve.
  • the associating of the annotation label information with the closed curve in S5 includes:
  • each of the thyroid disease type and the breast disease type has at least one disease condition label, and each disease condition label has a label color for representing the disease condition.
  • the condition label of the label information is extracted, and the label color of the disease label is obtained and set as the label color.
  • S512 Add the label color to the area delineated by the closed curve, and associate the label information with the closed curve.
  • the superimposition method is to superimpose on the original image in the form of an overlay (a virtualization technology mode that superimposes on the basis of the original image), which is convenient for the doctor to compare and observe.
  • the doctor can adjust the current result at any time until it is satisfied; further, output the annotation pixels in the annotation stack to the user end, and overlay them on the target image in a superimposed manner to achieve the area delineated by the closed curve
  • Annotated color is assigned inside; and then the binary image is converted into annotated image.
  • converting the binary image into an annotated image includes:
  • the effect of associating the label information with the binary image is achieved by summarizing and packaging the binary image with the label color in the area delineated by the closed curve and the label information of the label color to form the label image.
  • Annotated images are stored or transmitted in the form of data packets.
  • the storing the annotated image in the database in S5 includes:
  • S522 Detect whether the subtag of the pathology tag in the label information is complete; if the subtag is complete, generate a check success signal; if the subtag is incomplete, generate a prompt warning and output it to the user end.
  • the pathology label is a thyroid pathology label
  • Label, edge label and full icon annotation label information if yes, a check success signal is generated; if not, a prompt warning is generated and sent to the client to display it in the form of a pop-up window or a pop-up layer;
  • the subtags of the pathology label also have a breast disease type label, a BI-RADS label, a benign and malignant label, an echo pattern label, an edge label, an orientation label, and a shape label. ; If it is, a check success signal is generated; if not, a prompt warning is generated and sent to the client to display it in the form of a pop-up window or a pop-up layer.
  • the neutron label of the pathology label is completely selected, the comprehensiveness of the pathology label is ensured, and when faced with a situation that crosses departments or departments, it is convenient for other doctors to quickly and accurately understand the pathological details of the image, and the misdiagnosis rate is reduced. ; At the same time, by outputting a prompt warning to the user terminal, the user is prompted to indicate that the label of the detailed pathological information in the image is not complete.
  • the file information of the annotated image is obtained, the third-level node matching the file information is obtained through a query tree, the annotated image is saved in the third-level node, and the annotated image and its The target image is stored in the same third node to achieve the technical effect of associating the annotated image with the target image.
  • the label dialog box of step S502 is sent to the user terminal according to the prompt warning, so that the user can modify or supplement the sub-labels on the user terminal to obtain correct and complete label information.
  • an image labeling management apparatus 1 of this embodiment includes:
  • the creation module 11 is used to create a database for storing at least one image and having a query tree, the query tree is associated with file information of the image; wherein the image is a medical image, and the file information is a description image Attribute data
  • the retrieval module 12 obtains the display request sent by the client, obtains the file information matching the display request from the query tree, retrieves the target image corresponding to the file information, and compares the target image and its The file information is sent to the client for display;
  • the binary module 13 is used to obtain the information of the selected area selected by the user terminal on the target image, and send a binary dialog box to the user terminal to obtain binary data, and perform processing on the selected area information according to the binary data. Binarization processing obtains a binary image, and sends it to the user terminal for display; wherein the selected area information is a range pulled or delineated by the user terminal on the target image;
  • the growth module 14 is used to obtain the seed pixels and similarity thresholds selected by the user terminal on the binary image, perform region growth processing on the binary image to obtain a closed curve, and send the closed curve to the user end;
  • the annotation module 15 is used to obtain the annotation label information sent by the client, and associate the annotation label information with the closed curve, convert the binary image into an annotation image, and store the annotation image in the The database; wherein the area delineated by the closed curve is a lesion area in the target image, and the label information is feature data describing the characteristics of the lesion area.
  • This technical solution is based on image detection technology in the field of artificial intelligence.
  • the selected area information is determined in the target image, and the selected area information is binarized to obtain a binary image, Receive the seed coordinate data output by the user terminal, perform region growth processing on the binary image according to the seed coordinate data and the similarity threshold to achieve region extraction, and use the boundary pixels of the extracted region as a closed curve to complete the image processing.
  • the present application also provides a computer system including a plurality of computer devices 5.
  • the components of the image labeling management apparatus 1 of the second embodiment can be dispersed in different computer devices, and the computer devices can be executed Program smart phones, tablet computers, notebook computers, desktop computers, rack servers, blade servers, tower servers or cabinet servers (including independent servers, or server clusters composed of multiple servers), etc.
  • the computer device in this embodiment at least includes but is not limited to: a memory 51 and a processor 52 that can be communicatively connected to each other through a system bus, as shown in FIG. 10. It should be pointed out that FIG. 10 only shows a computer device with components, but it should be understood that it is not required to implement all the illustrated components, and more or fewer components may be implemented instead.
  • the memory 51 (ie, readable storage medium) includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), Read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable read only memory (PROM), magnetic memory, magnetic disk, optical disk, etc.
  • the memory 51 may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device.
  • the memory 51 may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), and a Secure Digital (SD).
  • SD Secure Digital
  • the memory 51 may also include both the internal storage unit of the computer device and its external storage device.
  • the memory 51 is generally used to store an operating system and various application software installed in a computer device, such as the program code of the image labeling management apparatus in the first embodiment, and the like.
  • the memory 51 can also be used to temporarily store various types of data that have been output or will be output.
  • the processor 52 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips.
  • the processor 52 is generally used to control the overall operation of the computer equipment.
  • the processor 52 is configured to run the program code or processing data stored in the memory 51, for example, to run an image labeling management device, so as to implement the image labeling management method of the first embodiment.
  • the computer-readable storage medium may be non-volatile or volatile, and includes multiple storage media, such as flash memory, hard disk, and multimedia.
  • the computer-readable storage medium of this embodiment is used to store the image labeling management device, and when executed by the processor 52, the image labeling management method of the first embodiment is implemented.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Radiology & Medical Imaging (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Epidemiology (AREA)
  • Quality & Reliability (AREA)
  • Public Health (AREA)
  • Primary Health Care (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

一种图像标注管理方法、装置、计算机系统及可读存储介质，基于人工智能领域，方法包括以下步骤：创建用于储存至少一张图像并具有查询树的数据库；从查询树中获取与显示请求匹配的文件信息，并调取与文件信息对应的目标图像，将目标图像及其文件信息发送用户端显示；获取选定区域信息并向用户端发送二值对话框以获取二值数据，根据二值数据对选定区域信息进行二值化处理获得二值图像；获取种子像素及相似阈值，并对二值图像进行区域生长处理以获得闭合曲线；将标注标签信息与闭合曲线关联，使二值图像转为标注图像。本方法提高了病灶区域的标注效率，缩短了标注时间，便于其他科室医生能够快速准确获知该图像中的病灶区域和病灶信息。

Description

图像标注管理方法、装置、计算机系统及可读存储介质
本申请要求于2020年2月25日提交中国专利局、申请号为CN202010115057.1，发明名称为“图像标注管理方法、装置、计算机系统及可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，其涉及到人工智能的计算机视觉技术，尤其涉及一种图像标注管理方法、装置、计算机系统及可读存储介质。
背景技术
从2019全国癌症统计报告可知，乳腺癌和甲状腺癌是排名前八的常见肿瘤，且发病率有逐年上升的趋势，形势十分严峻。然而，乳腺癌和甲状腺是一种潜在的可治愈的疾病，临床实践证明，早发现、早诊断、早治疗是降低乳腺癌死亡率的关键。如今，在治疗方面，利用影像学的手段，对乳腺、甲状腺等进行临床诊断比较普遍，其中超声影像检查以其检查无创性、影像实时性、无电离辐射、价格较低等诸多优点受到广泛应用，近年来，随着人们健康意识的不断提高，在浅表器官检查领域，临床医生和患者对超声检查的需求急剧增加；但由于其存在对比度较差、操作人员专业性要求较高以及超声科日常检查工作量较大的问题，对工作效率和准确性都提出了更高的要求。与此同时，也存在着医疗资源分布不均、基层设施较差以及专业医疗人员匮乏等一系列问题，很难快速实现病灶的检出和良恶性判断。
为了解决上述问题，市面上出现了越来越多的浅表超声影像AI产品可以辅助医生进行诊断，该产品需要专业医生对浅表超声(甲状腺超声、乳腺超声)影像病灶进行标注，然而发明人意识到当前对浅表超声病灶的标注通常采用手工标注的方式完成，这种方式不仅标注效率低，时间长，而且准确度也会因标注者的经验差异而不同，导致病灶标注质量参差不齐；又由于当前的数据库仅用于储存超声图像，未对图像中病灶区域进行圈定和标注，因此不便于其他科室的医生快速准确了解该图像中的病灶区域和病灶信息，为医生诊断带来了极大的不便。
发明内容
本申请的目的是提供一种图像标注管理方法、装置、计算机系统及可读存储介质，用于解决现有技术存在的标注效率低、时间长、标注质量参差不齐，以及对超声图像缺乏有效的管理，不便于其他科室的医生快速准确了解该图像中的病灶区域和病灶信息的问题。
为实现上述目的，本申请提供一种图像标注管理方法，包括以下步骤：
创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
为实现上述目的，本申请还提供一种图像标注管理装置，包括：
创建模块，用于创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
调取模块，获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
二值模块，用于获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
生长模块，用于获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
标注模块，用于获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
为实现上述目的，本申请还提供一种计算机系统，其包括多个计算机设备，各计算机设备包括存储器、处理器以及存储在存储器上并可在处理器上运行的计算机程序，所述多个计算机设备的处理器执行所述计算机程序时共同实现上述图像标注管理方法的步骤。
为实现上述目的，本申请还提供一种计算机可读存储介质，其包括多个存储介质，各存储介质上存储有计算机程序，所述多个存储介质存储的所述计算机程序被处理器执行时共同实现上述图像标注管理方法的步骤。
本申请提供的图像标注管理方法、装置、计算机系统及可读存储介质，通过创建用于储存图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；以对图像进行有效管理，消除了当前管理系统对图像管理混乱的情况；通过查询树并根据所述显示请求从数据库中调取目标图像及其文件信息，并通过用户端显示所述目标图像及其文件信息；通过二值数据对选定区域进行二值化处理，以获得用户降低无关的图像信息对医生的干扰；通过种子像素和相似阈值对二值图像进行区域生长处理，以获得闭合曲线，所述闭合曲线所圈定的区域即为医生需要标注的病灶区域，实现了自动且精准绘制闭合曲线的目的，提高了绘制效率，缩短了绘制时间，同时还保证了闭合曲线的绘制标注质量；通过标注标签对所述闭合曲线圈定的区域赋以颜色，以对该区域进行颜色标注，以便于其他科室部门的医生仅需通过闭合曲线及颜色即可快速获知病灶区域的范围及其病理特征，提高了科室部门间信息沟通效率。
附图说明
图1为本申请图像标注管理方法实施例一的流程图；
图2示意性示出了根据本申请实施例一的图像标注管理方法的环境应用示意图；
图3为本申请图像标注管理方法实施例一的步骤S3中向用户端发送二值对话框以获取二值数据的流程图；
图4为本申请图像标注管理方法实施例一的步骤S3中根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像的流程图；
图5为本申请图像标注管理方法实施例一的步骤S4中种子像素及相似阈值的获得方式的流程图；
图6为本申请图像标注管理方法实施例一的步骤S4中对二值图像进行区域生长处理以获得闭合曲线的流程图；
图7为本申请图像标注管理方法实施例一的步骤S5中将所述标注标签信息与所述闭合 曲线关联的流程图；
图8为本申请图像标注管理方法实施例一的步骤S5中将所述标注图像储存至所述数据库的流程图；
图9为本申请图像标注管理装置实施例二的程序模块示意图；
图10为本申请计算机系统实施例三中计算机设备的硬件结构示意图。
附图标记：
1、图像标注管理装置 2、服务器 3、网络 4、用户端
5、计算机设备 11、创建模块 12、调取模块 13、二值模块
14、生长模块 15、标注模块 51、存储器 52、处理器
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
本申请提供的图像标注管理方法、装置、计算机系统及可读存储介质，适用于计算机领域，为提供一种基于创建模块、调取模块、二值模块、生长模块和标注模块的图像标注管理方法。本申请通过创建用于储存图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；通过查询树并根据所述显示请求从数据库中调取目标图像及其文件信息，并通过用户端显示所述目标图像及其文件信息；通过二值数据对选定区域进行二值化处理；通过种子像素和相似阈值对二值图像进行区域生长处理，以获得闭合曲线，所述闭合曲线所圈定的区域即为医生需要标注的病灶区域；通过标注标签对所述闭合曲线圈定的区域赋以颜色，以对该区域进行颜色标注。
实施例一
请参阅图1，本实施例的一种图像标注管理方法，包括以下步骤：
S1：创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
S2：获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
S3：获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
S4：获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
S5：获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
图2示意性示出了根据本申请实施例一的图像标注管理方法的环境应用示意图。
在示例性的实施例中，图像标注管理方法运行在服务器2中，并在服务器2中创建用于储存图像并具有查询树的数据库；服务器2通过网络3与多个用户端4连接。
所述查询树包括第一级节点、第二级节点和第三级节点；所文件信息包括一级信息和二级信息，所述一级信息是描述图像所有者数据的信息，所述二级信息是基于一级信息描述图像基本情况的信息；较佳地，所述一级信息为病人信息，所述二级信息为基于所述病人信息的检查时间；例如，所述病人信息可包括但不限于：姓名，年龄，身份证号码；其 中，所述检查时间为病人拍摄所述图像的时间，可以年/月/日的形式展示。将所述一级信息储存在第一级节点中，将所述二级信息储存与所述一级信息对应的第二级节点中，将所述图像储存在与所述二级信息对应的第三级节点中，实现所述查询树与文件信息的关联；
需要说明的是，所述查询树为基于QTreeWidget创制的文件树列表，按照三级节点展示所述文件信息；其中，QTreeWidget是一种树形的部件，它以树的形式显示各个项，它的每个项使用QTreeWidgetItem来表示。所述对话框为基于QDialog所制成，其中，QDialog类是对话框窗口的基类，对话框窗口是主要用于短期任务以及和用户进行简要通讯的顶级窗口，QDialog可以是模式的也可以是非模式的，QDialog支持扩展性并且可以提供返回值。进一步地，在所述第二级节点下储存的检查时间按照检查时间升序或降序排列，方便医生进行选择和标注。
所述显示请求包括用于描述图像所有者信息的第一请求和用于描述图像基本情况的第二请求；服务器2根据显示请求在所述查询树的第一级节点中获取与所述第一请求匹配的一级信息，在所述一级信息下的第二级节点中获取与所述第二请求匹配的二级信息，提取所述二级信息下的第三级节点中储存的图像，并将其设为目标图像；将所述目标图像和目标图像的文件信息发送用户端显示。较佳地，所述第三级节点中还储存有诊断信息，所述诊断信息是专业医生对所述图像病灶所作出的诊断结论，其为经结构化处理后的数据信息，可与目标图像一并发送用户端显示，以便于将其输出至用户端并在病人信息栏中显示；其中，所述文件信息和诊断信息在用户端上通过病人信息栏显示，以便于医生可以参考病理表信息进行标注，这样可以给标注医生提供除了图像之外更多的信息，方便医生作出更加精准的标注。进一步地，在将所述目标图像及其文件信息和诊断信息加载至所述用户端前，初始化所述病人信息栏使其恢复至空白状态，以便于显示所述文件信息和诊断信息。
用户端4可通过拖动鼠标在目标图像上拉取获得选定区域信息，服务器2识别所述选定区域信息并向用户端4发送二值对话框，用户可通过用户端4在二值对话框中输入二值数据，对服务器2的二值化处理操作进行参数调节；服务器2根据所述二值数据对选定区域信息进行二值化处理，获得二值图像并将其发送用户端4显示。由于通过选定区域信息获得二值化处理的范围，因此，仅需对需要的范围进行二值化处理，而无需对目标图像进行整体二值化处理。用户可通过用户端4在二值图像中点选某一像素点，或拉取具有至少一个像素点的矩形框，获得种子像素；服务器2根据种子像素向用户端4发送相似对话框，用户可通过用户端4在相似对话框中输入相似阈值，对服务器2的区域生长处理操作进行参数调节；服务器2根据所述相似阈值和种子像素，对二值图像进行区域生长处理获得相似区域，沿所述相似区域的边界绘制闭合曲线，并将其发送所述用户端4显示。因病灶区域的灰度值与正常区域的灰度值是不同的，因此，通过二值化处理排除与病灶区域灰度值区别较大的正常区域，通过区域生长处理排除与病灶区域灰度值区别较小的正常区域；而通过二值化处理有助于医生更准确的识别出种子像素并对其点选。服务器2根据用户端4触发的标签请求，向用户端4发送标签对话框，用户端4在标签对话框中选择病理标签，并将其作为标注标签信息发送服务器2，服务器2根据标注标签信息中的标签颜色对闭合曲线圈定的区域赋以颜色，此时，所述二值图像将转为标注图像，将所述标注图像储存至所述数据库。
服务器2可以通过一个或多个网络3提供服务，网络3可以包括各种网络设备，例如路由器，交换机，多路复用器，集线器，调制解调器，网桥，中继器，防火墙，代理设备和/或等等。网络3可以包括物理链路，例如同轴电缆链路，双绞线电缆链路，光纤链路，它们的组合和/或类似物。网络3可以包括无线链路，例如蜂窝链路，卫星链路，Wi-Fi链路和/或类似物。
服务器2可以由单个或多个计算机设备(如，服务器)组成。该单个或多个计算设备可以包括虚拟化计算实例。虚拟化计算实例可以包括虚拟机，诸如计算机系统的仿真，操作系统，服务器等。计算设备可以基于定义用于仿真的特定软件(例如，操作系统，专用 应用程序，服务器)的虚拟映像和/或其他数据来加载虚拟机。随着对不同类型的处理服务的需求改变，可以在一个或多个计算设备上加载和/或终止不同的虚拟机。可以实现管理程序以管理同一计算设备上的不同虚拟机的使用。
在示例性的实施例中，所述S2中从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像包括：
S201：提取所述显示请求中的第一请求，并从所述查询树的第一级节点中获取与所述第一请求匹配的一级信息，获取与该一级信息对应的第二级节点。
例如，所述显示请求的第一请求为“张三”，则从查询树的第一节点中获取与“张三”匹配的一级信息，如：张三或姓名张三，并获取查询树中“张三或姓名张三”下的第二级节点。
S202：提取所述显示请求中的第二请求，并从所述第二级节点中获取与所述第二请求匹配的二级信息，获取与所述二级信息对应的第三级节点，并将所述一级信息和二级信息设为文件信息。
例如：显示请求的第二请求为“2018-01-20”，则从查询树的第二节点中获取与“2018-01-20”匹配的二级信息，如：2018-01-20或2018年1月20日，并获取查询树中“2018-01-20或2018年1月20日”下的第三级节点，并将张三、2018-01-20作为文件信息。
S203：将所述第三级节点中储存的图像设为目标图像。
需要说明的是，所述数据库中的查询树，图像、文件信息和诊断信息分别基于DICOM(Digital Imaging and Communications in Medicine)即医学数字成像和通信储存，DICOM是医学图像和相关信息的国际标准(ISO 12052)，它定义了质量能满足临床需要的可用于数据交换的医学图像格式，可用于处理、存储、打印和传输医学影像信息；Pydicom是一个处理DICOM文件的纯Python软件包，它可以通过非常容易的“Pythonic”的方式来提取和修改DICOM数据，修改后的数据还会借此生成新的DICOM文件。
在示例性的实施例中，所述S3中选定区域信息为用户端通过光标在所述目标图像上圈定一选中区域，所生成的用于描述该区域内所有像素的信息，其至少包括各像素的颜色信息、灰度值信息和坐标信息。
在一个优选的实施例中，请参阅图3，所述S3中的向用户端发送二值对话框以获取二值数据包括：
S301：向用户端发送二值对话框；其中，所述二值对话框具有上限输入框、下限输入框；
S302：接收所述用户端通过填写二值对话框生成的二值数据；其中，所述二值数据是用户端在上限输入框和下限输入框输入的处理上限阈值和处理下限阈值；
本步骤中，所述二值对话框上的上限输入框和下限输入框可通过SliderBar(滑条控件)的方式设置，用户端可通过调节SliderBar获得处理上限阈值、处理下限阈值的以及相似阈值。
在一个优选的实施例中，请参阅图4，所述S3中的根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像包括：
S311：提取选定区域信息中所有像素的灰度值；
S312：将灰度值大于处理上限阈值的像素设为过暗像素，将所述过暗像素的灰度值调高，例如，将全黑像素的灰度值调节为255；
S313：将灰度值小于处理下限阈值的像素设为过亮像素，将所述过亮像素的灰度值调低，例如，将全白像素的灰度值调节为0。
因此，实现对选定区域信息进行降噪，以降低无关的图像信息对医生的干扰，便于医生准确识别二值图像中的有用信息。
需要说明的是，所述二值图像通过overlay(是一种在原图像基础上进行叠加的虚拟化 技术模式)的形式叠加在原图上，不仅方便医生进行对比观察，还无需对原图进行的修改，避免了原图因修改不可恢复的情况发生。
在一个优选的实施例中，请参阅图5，所述S4中种子像素及相似阈值可通过以下方式获得：
S401：接收用户端在二值图像上点击或圈定的像素，并将该像素设为种子像素；
示例性地，创建用于储存像素的种子堆栈，接收用户端通过在二值信息图像上点击或圈定所生成的坐标数据，将所述坐标数据所对应的像素设为种子像素，并将其储存至所述种子堆栈。
S402：根据所述种子像素向用户端发送相似对话框，并接收用户端在所述相似对话框中输入的相似阈值。
本步骤中，所述相似对话框上的相似输入框可通过SliderBar(滑条控件)的方式设置，用户端可通过调节SliderBar获得相似阈值。
在一个优选的实施例中，请参阅图6，所述S4中对二值图像进行区域生长处理以获得闭合曲线包括：
S411：提取所述种子像素的灰度值，根据所述灰度值和相似阈值获得相似区间；
示例性地，提取所述种子像素的灰度值，并将该灰度值设为计算值，其中，若所述种子像素为一个，则提取该种子像素的灰度值并将其设为计算值，若所述种子像素为两个或两个以上，则提取各种子像素的灰度值，将各灰度值相加并除以种子像素的个数，获得平均灰度值，将该平均灰度值设为计算值；
将所述计算值与相似阈值相加获得相似上限，将所述计算值与相似阈值相减获得相似下限，编制以相似上限和相似下限为上限值和下限值的相似区间。
S412：将所述二值图像中灰度值处于所述相似区间的像素设为相似像素，并将所述种子像素和相似像素储存至预设的种子堆栈中；
示例性地，提取所述二值图像中像素的灰度值，将灰度值属于相似区间的像素设为相似像素，并将其储存至所述种子堆栈中。
S413：提取所述种子堆栈中位于边界的像素并将其设为边界像素，沿所述边界像素在所述二值图像上绘制闭合曲线。
示例性地，所述步骤S413中沿所述边界像素在所述二值图像上绘制闭合曲线还包括：
将所述闭合曲线的各个像素及其坐标储存至线条堆栈；用户端通过光标在闭合曲线上进行点选，将所述光标点选的像素设为采样像素；在所述线条堆栈中，将所述采样像素两侧的线条像素设为调整像素，其中，可根据需要设置调整像素的数量；用户端的光标移动至所述目标图像中某一位置并点选，生成具有到位坐标信息的到位信号；接收由所述用户端输出的到位信号，将所述采样像素的坐标更改为所述到位信号的到位坐标信息，并将所述采样像素储存在所述线条堆栈中；计算所述起始坐标信息和到位坐标信息之间的直线距离，并根据所述直线距离和采样像素与调整像素的间隔计算获得移动距离，其中，按照衰减函数D(s)＝D0e^(-a(s+1))调整所述调整像素的移动坐标，其中，a为缩放因子，s为当前调整像素和采样像素之间的间距，D0为直线距离，D(s)为调整像素所移动的距离，例如，当前采样像素和调整像素之间的间距为3所述直线距离为10个像素值，则当前调整像素的移动距离为10e^(-4a)，其移动方向与采样像素的移动方向保持一致；根据采样像素的移动方向，按照所述移动距离改变调整像素的坐标，并将所述调整像素储存至线条堆栈中；将所述线条堆栈中的像素输出至显示设备，以实现调整所述闭合曲线在目标图像中位置的效果。
在示例性的实施例中，所述标注标签信息可通过以下方式获得：
S501：预创建用于储存标签对话框的选择数据库，所述标签对话框中具有病理标签。
示例性地，预先创建选择数据库用于储存标签对话框，根据用户需求在标签对话框中制定病理标签，例如，所述病理标签可包括：钙化、肿块、结构扭曲、不对称等用于描述 乳腺病灶的病理名称；通过在病理标签上设置选择框或勾选框，使其加载在所述标签对话框中，以便于用户通过点击选择框或勾选框，即可选择需要的病理标签。
进一步地，标签对话框包括甲状腺对话框和乳腺对话框；所述甲状腺对话框包括：甲状腺病种标签、甲状腺内部构成标签、内部强回声标签、回声类型标签、形状标签、边界标签、边缘标签、全图标注标签信息；所述标签对话框为乳腺对话框包括：乳腺病种标签、BI-RADS标签、良恶性标签、回声模式标签、边缘标签、方位标签、形状标签；则接收所述用户端在所述乳腺对话框中选择的病理标签的子标签。
S502：根据所述闭合曲线从所述选择数据库中提取标签对话框，并将其输出至用户端上，使其在用户端上弹出。
需要说明的是，所述标签对话框为基于QDialog制成的疾病选择预定义模板，其中，QDialog类是对话框窗口的基类，对话框窗口是主要用于短期任务以及和用户进行简要通讯的顶级窗口，QDialog可以是模式的也可以是非模式的，QDialog支持扩展性并且可以提供返回值；通过采用标签对话框避免了医生对病灶的描述因人而异的情况，因此，避免了其他医生因对标注医生的标注出现误解，造成误诊的情况。
S503：接收用户端在所述标签对话框中选择的病理标签以获得标注标签信息。
在示例性的实施例中，用户通过点击选择框或勾选框选择需要的病理标签的子标签，以获得标注标签信息，以实现医生对所述闭合曲线圈定的区域进行定义或标记；根据所述闭合曲线向用户端输出具有病种选项的种类对话框并使其在用户端弹出，所述病种选项包括甲状腺选项和乳腺选项；若所述用户端点选甲状腺选项，则向用户端输出甲状腺对话框；若所述用户端点选乳腺选项，则向用户端输出乳腺对话框；所述用户端可在所述甲状腺对话框或乳腺对话框中选择需要病理标签的子标签，以获得用于描述闭合曲线圈定区域的病灶特征的标注标签信息。
在一个优选的实施例中，请参阅图7，所述S5中将所述标注标签信息与所述闭合曲线关联包括：
S511：提取所述标注标签信息的标注颜色。
于本申请中，所述甲状腺病种和乳腺病种中分别具有至少一个病情标签，每个病情标签具有用于代表该病情的标注颜色。
故本步骤中，提取所述标注标签信息的病情标签，并获取该病情标签的标注颜色，将其设为标注颜色。
S512：在所述闭合曲线圈定的区域内赋以所述标注颜色，并将所述标注标签信息与闭合曲线关联。
示例性地，复制所述闭合曲线内圈定的像素作为标注像素，并将其储存在标注堆栈中；提取所述标注标签信息中病种标签的标注颜色，并根据标注颜色对所述标注像素的RGB分量赋值，使所述标注像素的颜色与所述标注颜色一致；其中，通过将标注堆栈中的标注像素发送至用户端，以在用户端的闭合曲线圈定区域内赋以标注颜色。
需要说明的是，所述叠加的方式为通过overlay(是一种在原图像基础上进行叠加的虚拟化技术模式)的形式叠加在原图上，方便医生进行对比观察。医生可以随时调整当前结果，直至满意为止；进一步地，将所述标注堆栈中的标注像素输出至所述用户端，并以叠加的方式覆盖在所述目标图像上，实现对闭合曲线圈定的区域内赋以标注颜色；进而使二值图像转为标注图像。
进一步地，使所述二值图像转为标注图像包括：
通过将在闭合曲线圈定的区域内赋有标注颜色的二值图像，和所述标注颜色的标注标签信息汇总并打包形成标注图像，实现将标注标签信息和二值图像关联的效果，其中，所述标注图像以数据包形式储存或传输。
在一个优选的的实施例中，请参阅图8，所述S5中将所述标注图像储存至所述数据库包括：
S521：提取所述标注图像中的标注标签信息。
S522：检测所述标注标签信息中病理标签的子标签是否完整；若所述子标签完整，则生成检查成功信号；若所述子标签不完整，则生成提示警告并将其输出至所述用户端。
示例性地，若所述病理标签为甲状腺病理标签，则检查所述病例标签的子标签中是否同时具有甲状腺病种标签、甲状腺内部构成标签、内部强回声标签、回声类型标签、形状标签、边界标签、边缘标签和全图标注标签信息，若是，则生成检查成功信号；若否，则生成提示警告，并将其发送用户端以弹窗或弹出层的形式显示；
若所述病理详细信息中具有乳腺选项，则检查所述病理标签的子标签中是否同时具有乳腺病种标签、BI-RADS标签、良恶性标签、回声模式标签、边缘标签、方位标签和形状标签；若是，则生成检查成功信号；若否，则生成提示警告，并将其发送用户端以弹窗或弹出层的形式显示。
因此，通过检查病理标签中子标签是否完全选定，保证了病理标签的全面性，在面临跨部门或科室的情况时，便于其他医生能够快速准确了解该图像的病理详细信息，降低了误诊率；同时，通过向用户端输出提示警告，以向使用者提示标注图像中病理详细信息的标签不全的情况。
S523：根据检查成功信号将所述标注图像保存至数据库。
示例性地，获得所述标注图像的文件信息，通过查询树获得与所述文件信息匹配的第三级节点，将所述标注图像保存至所述第三级节点中，通过将标注图像及其目标图像储存在同一第三节点中，实现将标注图像和目标图像关联的技术效果。根据提示警告向用户端发送所述步骤S502的标签对话框，以便于用户在用户端上修改或补充子标签，以获得正确完整的标注标签信息。
实施例二
请参阅图9，本实施例的一种图像标注管理装置1，包括：
创建模块11，用于创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
调取模块12，获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
二值模块13，用于获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
生长模块14，用于获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
标注模块15，用于获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
本技术方案基于人工智能领域的图像检测技术，根据所述绘制信息中的选中信息，在目标图像中确定选定区域信息，并对该选定区域信息进行二值化处理以获得二值图像，接收用户端输出的种子坐标数据，根据所述种子坐标数据和相似阈值对二值图像进行区域生长处理以实现区域提取，将提取的区域的边界像素作为闭合曲线以完成图像的处理。
实施例三
为实现上述目的，本申请还提供一种计算机系统，该计算机系统包括多个计算机设备5，实施例二的图像标注管理装置1的组成部分可分散于不同的计算机设备中，计算机设备可 以是执行程序的智能手机、平板电脑、笔记本电脑、台式计算机、机架式服务器、刀片式服务器、塔式服务器或机柜式服务器(包括独立的服务器，或者多个服务器所组成的服务器集群)等。本实施例的计算机设备至少包括但不限于：可通过系统总线相互通信连接的存储器51、处理器52，如图10所示。需要指出的是，图10仅示出了具有组件-的计算机设备，但是应理解的是，并不要求实施所有示出的组件，可以替代的实施更多或者更少的组件。
本实施例中，存储器51(即可读存储介质)包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘等。在一些实施例中，存储器51可以是计算机设备的内部存储单元，例如该计算机设备的硬盘或内存。在另一些实施例中，存储器51也可以是计算机设备的外部存储设备，例如该计算机设备上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(Flash Card)等。当然，存储器51还可以既包括计算机设备的内部存储单元也包括其外部存储设备。本实施例中，存储器51通常用于存储安装于计算机设备的操作系统和各类应用软件，例如实施例一的图像标注管理装置的程序代码等。此外，存储器51还可以用于暂时地存储已经输出或者将要输出的各类数据。
处理器52在一些实施例中可以是中央处理器(Central Processing Unit，CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器52通常用于控制计算机设备的总体操作。本实施例中，处理器52用于运行存储器51中存储的程序代码或者处理数据，例如运行图像标注管理装置，以实现实施例一的图像标注管理方法。
实施例四
为实现上述目的，本申请还提供一种计算机可读存储系统，所述计算机可读存储介质可以是非易失性，也可以是易失性，其包括多个存储介质，如闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘、服务器、App应用商城等等，其上存储有计算机程序，程序被处理器52执行时实现相应功能。本实施例的计算机可读存储介质用于存储图像标注管理装置，被处理器52执行时实现实施例一的图像标注管理方法。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种图像标注管理方法，其中，包括以下步骤：
    创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
    获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
    获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
    获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
    获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
  2. 根据权利要求1所述的图像标注管理方法，其中，所述向用户端发送二值对话框以获取二值数据包括：
    向用户端发送二值对话框；其中，所述二值对话框具有上限输入框、下限输入框；
    接收所述用户端通过填写二值对话框生成的二值数据；其中，所述二值数据是用户端在上限输入框和下限输入框输入的处理上限阈值和处理下限阈值。
  3. 根据权利要求1所述的图像标注管理方法，其中，所述根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像包括：
    提取选定区域信息中所有像素的灰度值；
    将灰度值大于处理上限阈值的像素设为过暗像素，将所述过暗像素的灰度值调高；
    将灰度值小于处理下限阈值的像素设为过亮像素，将所述过亮像素的灰度值调低。
  4. 根据权利要求1所述的图像标注管理方法，其中，所述种子像素及相似阈值可通过以下方式获得：
    接收用户端在二值图像上点击或圈定的像素，并将该像素设为种子像素；
    根据所述种子像素向用户端发送相似对话框，并接收用户端在所述相似对话框中输入的相似阈值。
  5. 根据权利要求1所述的图像标注管理方法，其中，所述对所述二值图像进行区域生长处理以获得闭合曲线包括：
    提取所述种子像素的灰度值，根据所述灰度值和相似阈值获得相似区间；
    将所述二值图像中灰度值处于所述相似区间的像素设为相似像素，并将所述种子像素和相似像素储存至预设的种子堆栈中；
    提取所述种子堆栈中位于边界的像素并将其设为边界像素，沿所述边界像素在所述二值图像上绘制闭合曲线。
  6. 根据权利要求1所述的图像标注管理方法，其中，所述将所述标注标签信息与所述闭合曲线关联包括：
    提取所述标注标签信息的标注颜色；
    在所述闭合曲线圈定的区域内赋以所述标注颜色，使标注标签信息与闭合曲线关联。
  7. 根据权利要求1所述的图像标注管理方法，其中，所述将所述标注图像储存至 所述数据库包括：
    提取所述标注图像中的标注标签信息；
    检测所述标注标签信息中病理标签的子标签是否完整；若所述子标签完整，则生成检查成功信号；若所述子标签不完整，则生成提示警告并将其输出至所述用户端；
    根据检查成功信号将所述标注图像保存至数据库。
  8. 一种图像标注管理装置，其中，包括：
    创建模块，用于创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
    调取模块，获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
    二值模块，用于获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
    生长模块，用于获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
    标注模块，用于获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
  9. 一种计算机系统，其包括多个计算机设备，各计算机设备包括存储器、处理器以及存储在存储器上并可在处理器上运行的计算机程序，其中，所述多个计算机设备的处理器执行所述计算机程序时共同实现所述图像标注管理方法，其包括以下步骤：
    创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
    获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
    获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
    获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
    获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
  10. 根据权利要求9所述的计算机系统，其中，所述向用户端发送二值对话框以获取二值数据包括：
    向用户端发送二值对话框；其中，所述二值对话框具有上限输入框、下限输入框；
    接收所述用户端通过填写二值对话框生成的二值数据；其中，所述二值数据是用户端在上限输入框和下限输入框输入的处理上限阈值和处理下限阈值。
  11. 根据权利要求9所述的计算机系统，其中，所述根据所述二值数据对所述选定 区域信息进行二值化处理获得二值图像包括：
    提取选定区域信息中所有像素的灰度值；
    将灰度值大于处理上限阈值的像素设为过暗像素，将所述过暗像素的灰度值调高；
    将灰度值小于处理下限阈值的像素设为过亮像素，将所述过亮像素的灰度值调低。
  12. 根据权利要求9所述的计算机系统，其中，所述种子像素及相似阈值可通过以下方式获得：
    接收用户端在二值图像上点击或圈定的像素，并将该像素设为种子像素；
    根据所述种子像素向用户端发送相似对话框，并接收用户端在所述相似对话框中输入的相似阈值。
  13. 根据权利要求9所述的计算机系统，其中，所述对所述二值图像进行区域生长处理以获得闭合曲线包括：
    提取所述种子像素的灰度值，根据所述灰度值和相似阈值获得相似区间；
    将所述二值图像中灰度值处于所述相似区间的像素设为相似像素，并将所述种子像素和相似像素储存至预设的种子堆栈中；
    提取所述种子堆栈中位于边界的像素并将其设为边界像素，沿所述边界像素在所述二值图像上绘制闭合曲线。
  14. 根据权利要求9所述的计算机系统，其中，所述将所述标注标签信息与所述闭合曲线关联包括：
    提取所述标注标签信息的标注颜色；
    在所述闭合曲线圈定的区域内赋以所述标注颜色，使标注标签信息与闭合曲线关联；
    所述将所述标注图像储存至所述数据库包括：
    提取所述标注图像中的标注标签信息；
    检测所述标注标签信息中病理标签的子标签是否完整；若所述子标签完整，则生成检查成功信号；若所述子标签不完整，则生成提示警告并将其输出至所述用户端；
    根据检查成功信号将所述标注图像保存至数据库。
  15. 一种计算机可读存储介质，其包括多个存储介质，各存储介质上存储有计算机程序，其中，所述多个存储介质存储的所述计算机程序被处理器执行时共同实现所述图像标注管理方法，其包括以下步骤：
    创建用于储存至少一张图像并具有查询树的数据库，所述查询树与所述图像的文件信息关联；其中，所述图像为医学图像，所述文件信息是描述图像属性的数据；
    获取用户端发送的显示请求，从所述查询树中获取与所述显示请求匹配的文件信息，并调取与所述文件信息对应的目标图像，将所述目标图像及其文件信息发送所述用户端显示；
    获取用户端在目标图像上选中的选定区域信息，并向用户端发送二值对话框以获取二值数据，根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像，并将其发送所述用户端显示；其中，所述选定区域信息是用户端在目标图像上拉取或圈定的范围；
    获取用户端在所述二值图像上选择的种子像素及相似阈值，并对所述二值图像进行区域生长处理以获得闭合曲线，将所述闭合曲线发送至所述用户端；
    获取所述用户端发送的标注标签信息，并将所述标注标签信息与所述闭合曲线关联，使所述二值图像转为标注图像，将所述标注图像储存至所述数据库；其中，所述闭合曲线圈定的区域是目标图像中的病灶区域，所述标注标签信息是描述所述病灶区域特征的特征数据。
  16. 根据权利要求15所述的计算机可读存储介质，其中，所述向用户端发送二值对话框以获取二值数据包括：
    向用户端发送二值对话框；其中，所述二值对话框具有上限输入框、下限输入框；
    接收所述用户端通过填写二值对话框生成的二值数据；其中，所述二值数据是用户端在上限输入框和下限输入框输入的处理上限阈值和处理下限阈值。
  17. 根据权利要求15所述的计算机可读存储介质，其中，所述根据所述二值数据对所述选定区域信息进行二值化处理获得二值图像包括：
    提取选定区域信息中所有像素的灰度值；
    将灰度值大于处理上限阈值的像素设为过暗像素，将所述过暗像素的灰度值调高；
    将灰度值小于处理下限阈值的像素设为过亮像素，将所述过亮像素的灰度值调低。
  18. 根据权利要求15所述的计算机可读存储介质，其中，所述种子像素及相似阈值可通过以下方式获得：
    接收用户端在二值图像上点击或圈定的像素，并将该像素设为种子像素；
    根据所述种子像素向用户端发送相似对话框，并接收用户端在所述相似对话框中输入的相似阈值。
  19. 根据权利要求15所述的计算机可读存储介质，其中，所述对所述二值图像进行区域生长处理以获得闭合曲线包括：
    提取所述种子像素的灰度值，根据所述灰度值和相似阈值获得相似区间；
    将所述二值图像中灰度值处于所述相似区间的像素设为相似像素，并将所述种子像素和相似像素储存至预设的种子堆栈中；
    提取所述种子堆栈中位于边界的像素并将其设为边界像素，沿所述边界像素在所述二值图像上绘制闭合曲线。
  20. 根据权利要求1所述的计算机可读存储介质，其中，所述将所述标注标签信息与所述闭合曲线关联包括：
    提取所述标注标签信息的标注颜色；
    在所述闭合曲线圈定的区域内赋以所述标注颜色，使标注标签信息与闭合曲线关联；
    所述将所述标注图像储存至所述数据库包括：
    提取所述标注图像中的标注标签信息；
    检测所述标注标签信息中病理标签的子标签是否完整；若所述子标签完整，则生成检查成功信号；若所述子标签不完整，则生成提示警告并将其输出至所述用户端；
    根据检查成功信号将所述标注图像保存至数据库。
PCT/CN2020/099291 2020-02-25 2020-06-30 图像标注管理方法、装置、计算机系统及可读存储介质 WO2021169122A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010115057.1A CN111324761B (zh) 2020-02-25 2020-02-25 图像标注管理方法、装置、计算机系统及可读存储介质
CN202010115057.1 2020-02-25

Publications (1)

Publication Number Publication Date
WO2021169122A1 true WO2021169122A1 (zh) 2021-09-02

Family

ID=71172900

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/099291 WO2021169122A1 (zh) 2020-02-25 2020-06-30 图像标注管理方法、装置、计算机系统及可读存储介质

Country Status (2)

Country Link
CN (1) CN111324761B (zh)
WO (1) WO2021169122A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113901933A (zh) * 2021-10-14 2022-01-07 中国平安人寿保险股份有限公司 基于人工智能的电子发票信息抽取方法、装置及设备
CN115274093A (zh) * 2022-07-26 2022-11-01 华东师范大学 生成包含自动标注文件的基准病理数据集的方法及系统

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111324761B (zh) * 2020-02-25 2023-10-13 平安科技（深圳）有限公司 图像标注管理方法、装置、计算机系统及可读存储介质
CN111914822B (zh) * 2020-07-23 2023-11-17 腾讯科技（深圳）有限公司 文本图像标注方法、装置、计算机可读存储介质及设备
CN112035774A (zh) * 2020-09-01 2020-12-04 平安付科技服务有限公司 网络页面生成方法、装置、计算机设备及可读存储介质
CN113689937A (zh) * 2021-07-07 2021-11-23 阿里巴巴新加坡控股有限公司 图像标注方法、存储介质和处理器
CN113806573A (zh) * 2021-09-15 2021-12-17 上海商汤科技开发有限公司 标注方法、装置、电子设备、服务器及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104462738A (zh) * 2013-09-24 2015-03-25 西门子公司 一种标注医学图像的方法、装置和系统
CN105608319A (zh) * 2015-12-21 2016-05-25 江苏康克移软软件有限公司 一种数字病理切片的标注方法及标注装置
CN106934228A (zh) * 2017-03-06 2017-07-07 杭州健培科技有限公司 基于机器学习的肺部气胸ct影像分类诊断方法
US20190228524A1 (en) * 2018-01-23 2019-07-25 Beijing Curacloud Technology Co., Ltd. System and method for medical image management
CN111324761A (zh) * 2020-02-25 2020-06-23 平安科技（深圳）有限公司 图像标注管理方法、装置、计算机系统及可读存储介质名称

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106780497B (zh) * 2016-11-28 2019-09-24 重庆大学 一种基于统计信息的器官血管树自动提取方法
CN109522898A (zh) * 2018-09-18 2019-03-26 平安科技（深圳）有限公司 手写样本图片标注方法、装置、计算机设备及存储介质
CN109978894A (zh) * 2019-03-26 2019-07-05 成都迭迦科技有限公司 一种基于三维乳腺彩超的病变区域标注方法及系统
CN110675940A (zh) * 2019-08-01 2020-01-10 平安科技（深圳）有限公司 病理图像标注方法、装置、计算机设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104462738A (zh) * 2013-09-24 2015-03-25 西门子公司 一种标注医学图像的方法、装置和系统
CN105608319A (zh) * 2015-12-21 2016-05-25 江苏康克移软软件有限公司 一种数字病理切片的标注方法及标注装置
CN106934228A (zh) * 2017-03-06 2017-07-07 杭州健培科技有限公司 基于机器学习的肺部气胸ct影像分类诊断方法
US20190228524A1 (en) * 2018-01-23 2019-07-25 Beijing Curacloud Technology Co., Ltd. System and method for medical image management
CN111324761A (zh) * 2020-02-25 2020-06-23 平安科技（深圳）有限公司 图像标注管理方法、装置、计算机系统及可读存储介质名称

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113901933A (zh) * 2021-10-14 2022-01-07 中国平安人寿保险股份有限公司 基于人工智能的电子发票信息抽取方法、装置及设备
CN115274093A (zh) * 2022-07-26 2022-11-01 华东师范大学 生成包含自动标注文件的基准病理数据集的方法及系统
CN115274093B (zh) * 2022-07-26 2023-06-23 华东师范大学 生成包含自动标注文件的基准病理数据集的方法及系统

Also Published As

Publication number Publication date
CN111324761A (zh) 2020-06-23
CN111324761B (zh) 2023-10-13

Similar Documents

Publication Publication Date Title
WO2021169122A1 (zh) 图像标注管理方法、装置、计算机系统及可读存储介质
US20200334810A1 (en) Methods and systems for dynamically training and applying neural network analyses to medical images
CA3017647C (en) Optical character recognition in structured documents
US20220207231A1 (en) Methods and systems for guided information viewing and storage platforms
WO2021073157A1 (zh) 图像管理显示方法、装置、计算机设备及存储介质
US6819785B1 (en) Image reporting method and system
US10789712B2 (en) Method and system for image analysis to detect cancer
US20170277663A1 (en) Digital content conversion and publishing system
US20200226174A1 (en) Cloud-based large-scale pathological image collaborative annotation method and system
CN110197715B (zh) 一种用于读片教学的医学影像浏览系统
US20160026858A1 (en) Image based search to identify objects in documents
WO2023015935A1 (zh) 一种体检项目推荐方法、装置、设备及介质
US11921795B2 (en) Data normalization and extraction system
CN111986194A (zh) 医学标注图像检测方法、装置、电子设备及存储介质
CA3018437A1 (en) Optical character recognition utilizing hashed templates
US20230147471A1 (en) Systems and methods to process electronic images to determine salient information in digital pathology
Wang et al. Technology standards in imaging: a practical overview
US20190042852A1 (en) Supplementing a media stream with additional information
US20230334663A1 (en) Development of medical imaging ai analysis algorithms leveraging image segmentation
CN116825269A (zh) 体检报告的处理方法、装置、电子设备和可读存储介质
WO2023024959A1 (zh) 图像标注方法、系统、设备和存储介质
CN113409280A (zh) 医学影像的处理方法、标注方法和电子设备
US20190065682A1 (en) Automatic generation of ui from annotation templates
EP4160606A1 (en) Exchange of data between an external data source and an integrated medical data display system
AU2018317910A1 (en) Processing data records and searching data structures that are stored in hardware memory and that are at least partly generated from the processed data records in generating an adaptive user interface

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20920999

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20920999

Country of ref document: EP

Kind code of ref document: A1