Image annotation method and system, device, storage medium, and program product
By generating and storing annotation subtasks to the target task queue, image annotation tasks are automatically executed, solving the problem that subtasks cannot be automatically chained together in existing technologies, and achieving efficient image annotation processing.
Patent Information
- Application Number
- PCT/CN2026/085717
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-09
- Filing Date
- 2026-03-24
- Publication Date
- 2026-11-12
AI Technical Summary
In existing image annotation processes, the sub-tasks cannot be automatically linked together, requiring a large amount of manual operation, resulting in low processing efficiency.
By receiving image annotation requests, generating and storing annotation subtasks to the target task queue, automatically executing each subtask in the execution order, and using the annotation model for annotation processing, the automatic chaining of subtasks is achieved.
It improves the processing efficiency of image annotation, reduces manual intervention, and enhances the accuracy and efficiency of annotation.
Smart Images

Figure CN2026085717_12112026_PF_FP_ABST
Abstract
Description
Image annotation methods, systems, devices, storage media, and program products
[0001] This application claims priority to Chinese Patent Application No. 202510599814.X, filed on May 9, 2025, entitled “Image Annotation Method, System, Device, Storage Medium and Program Product”, the entire contents of which are incorporated herein by reference. Technical Field
[0002] This application relates to the field of annotation technology, and in particular to an image annotation method, system, device, storage medium and program product. Background Technology
[0003] Image annotation technology has become a key step in building high-quality training datasets. In the field of autonomous driving, it is necessary to annotate elements such as road features, traffic lights, vehicles, and pedestrians.
[0004] In related technologies, a labeling request may include multiple sub-requests. After the labeling request is uploaded to the image labeling system, the image labeling system will determine the sub-task corresponding to each sub-request. Different sub-tasks are independent of each other. After the sub-task corresponding to each sub-request is completed, the next sub-task to be executed needs to be determined manually until all sub-tasks are completed. Then, the intermediate labeled images corresponding to each sub-task are selected manually and integrated to obtain the target labeled image.
[0005] However, the split subtasks cannot be automatically linked together, requiring a lot of manual operation to select the next subtask to be executed. The manual participation in linking the subtasks together makes the annotation process inefficient. Summary of the Invention
[0006] This application provides an image annotation method, system, device, storage medium, and program product to improve the processing efficiency of annotation.
[0007] In a first aspect, embodiments of this application provide an image annotation method, including:
[0008] Receive an image annotation request, the image annotation request is used to annotate a preset image, the image annotation request includes the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request;
[0009] Based on the configuration information corresponding to each annotation sub-request, generate annotation sub-tasks corresponding to each annotation sub-request, and store each annotation sub-task in the target task queue according to the execution order of each annotation sub-request;
[0010] Each annotation subtask is executed in the storage order of the target task queue to annotate the preset image and obtain the annotated image.
[0011] In one possible implementation, before storing each annotation subtask into the target task queue according to the execution order corresponding to each annotation sub-request, the method further includes:
[0012] Obtain at least one preset subtask and the execution order of the at least one preset subtask in the image annotation task. The preset subtask is a subtask in the image annotation task other than each annotation subtask.
[0013] Based on the execution order of each annotation sub-request, determine the sub-queue corresponding to each annotation sub-task;
[0014] The storage location of the sub-queue in the target task queue is determined according to the execution order of at least one of the preset sub-tasks in the image annotation task;
[0015] Each labeled subtask is stored in the target task queue according to the specified storage location.
[0016] In one possible implementation, generating a labeling sub-task corresponding to each labeling sub-request based on the configuration information corresponding to each labeling sub-request includes:
[0017] For any given annotation sub-request, the annotation object corresponding to the annotation sub-request and the annotation strategy corresponding to the annotation object are determined based on the configuration information.
[0018] For any given annotation sub-request, an annotation sub-task corresponding to the annotation sub-request is generated based on the annotation object and the annotation strategy corresponding to the annotation object.
[0019] In one possible implementation, for any given annotation subtask, each annotation subtask is executed to annotate the preset image to obtain the annotated image, including:
[0020] Obtain the annotation model corresponding to the annotation subtask from the annotation strategy corresponding to the annotation subtask;
[0021] The annotation subtask is executed by the annotation model corresponding to the annotation subtask to annotate the preset image and obtain the annotated image.
[0022] In one possible implementation, the target task queue is stored in an initial annotation queue; before the annotation subtasks are executed according to their storage order in the target task queue to annotate the preset image and obtain the annotated image, the method further includes:
[0023] Obtain the target priority corresponding to the target task queue;
[0024] If it is determined that the storage of the target task queue in the initial annotation queue does not meet the target priority requirements, then the storage location of the target task queue is updated in the initial annotation queue to obtain the target annotation queue. The target annotation queue includes multiple task queues to be executed, and the target task queue is included among the multiple task queues to be executed.
[0025] In one possible implementation, the method further includes:
[0026] The verification identifiers of each target subtask in the target task queue are detected;
[0027] If the verification identifier of a target subtask is a verification anomaly identifier, the target subtask is identified as an anomaly subtask;
[0028] The abnormal subtask is handled in an exception manner, which includes rolling back execution or skipping execution.
[0029] In one possible implementation, the method further includes:
[0030] Receive the task change request corresponding to the target task queue;
[0031] If the task change request is a request to change the priority of the target task queue, the task change request includes the change priority corresponding to the target task queue; according to the change priority, the storage location of the target task queue is updated in the target label queue;
[0032] If the task change request is a request to change the labeled subtasks of the target task queue, the task change request includes change information corresponding to the labeled subtasks; and the labeled subtasks of the target task queue are updated according to the change information.
[0033] Secondly, embodiments of this application provide an image annotation system, including: a request receiving module, a subtask generation module, a task queue generation module, and an annotation module, wherein...
[0034] The request receiving module is used to receive image annotation requests, which are used to annotate a preset image. The image annotation request includes the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request.
[0035] The subtask generation module is used to generate annotation subtasks corresponding to each annotation sub-request based on the configuration information corresponding to each annotation sub-request.
[0036] The task queue generation module is used to store each annotation subtask into the target task queue according to the execution order corresponding to each annotation sub-request;
[0037] The annotation module is used to execute each annotation subtask according to the storage order of each annotation subtask in the target task queue, so as to perform annotation processing on the preset image and obtain the annotated image.
[0038] In one possible implementation, the task queue generation module is further configured to:
[0039] Obtain at least one preset subtask and the execution order of the at least one preset subtask in the image annotation task. The preset subtask is a subtask in the image annotation task other than each annotation subtask.
[0040] Based on the execution order of each annotation sub-request, determine the sub-queue corresponding to each annotation sub-task;
[0041] The storage location of the sub-queue in the target task queue is determined according to the execution order of at least one of the preset sub-tasks in the image annotation task;
[0042] Each labeled subtask is stored in the target task queue according to the specified storage location.
[0043] In one possible implementation, the subtask generation module is specifically used for:
[0044] For any given annotation sub-request, the annotation object corresponding to the annotation sub-request and the annotation strategy corresponding to the annotation object are determined based on the configuration information.
[0045] For any given annotation sub-request, an annotation sub-task corresponding to the annotation sub-request is generated based on the annotation object and the annotation strategy corresponding to the annotation object.
[0046] In one possible implementation, the annotation module 704 is further configured to:
[0047] Obtain the annotation model corresponding to the annotation subtask from the annotation strategy corresponding to the annotation subtask;
[0048] The annotation subtask is executed by the annotation model corresponding to the annotation subtask to annotate the preset image and obtain the annotated image.
[0049] In one possible implementation, the storage module is used to store the target task queue into the initial labeling queue, and the storage module has the following functions:
[0050] Obtain the target priority corresponding to the target task queue;
[0051] If it is determined that the storage of the target task queue in the initial annotation queue does not meet the target priority requirements, then the storage location of the target task queue is updated in the initial annotation queue to obtain the target annotation queue. The target annotation queue includes multiple task queues to be executed, and the target task queue is included among the multiple task queues to be executed.
[0052] In one possible implementation, the detection module is used for:
[0053] The verification identifiers of each target subtask in the target task queue are detected;
[0054] If the verification identifier of a target subtask is a verification anomaly identifier, the target subtask is identified as an anomaly subtask;
[0055] The abnormal subtask is handled in an exception manner, which includes rolling back execution or skipping execution.
[0056] In one possible implementation, the change module is used for:
[0057] Receive the task change request corresponding to the target task queue;
[0058] If the task change request is a request to change the priority of the target task queue, the task change request includes the change priority corresponding to the target task queue; according to the change priority, the storage location of the target task queue is updated in the target label queue;
[0059] If the task change request is a request to change the labeled subtasks of the target task queue, the task change request includes change information corresponding to the labeled subtasks; and the labeled subtasks of the target task queue are updated according to the change information.
[0060] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;
[0061] The memory stores computer-executed instructions;
[0062] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.
[0063] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.
[0064] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.
[0065] The image annotation method, system, device, storage medium, and program product provided in this application embodiment allow an image annotation request to include the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request. Based on the configuration information corresponding to each annotation sub-request, annotation sub-tasks corresponding to each annotation sub-request can be generated, and each annotation sub-task can be stored in a target task queue according to the execution order of each annotation sub-request. Each annotation sub-task is executed according to its storage order in the target task queue to annotate a preset image, obtaining an annotated image. Annotation tasks can be automatically executed through the target task queue to obtain an annotated image, eliminating the need for manual concatenation between sub-tasks and improving processing efficiency. Attached Figure Description
[0066] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.
[0067] Figure 1 is a schematic diagram of an application scenario provided by an embodiment of this application;
[0068] Figure 2 is a flowchart illustrating an image annotation method provided in an embodiment of this application;
[0069] Figure 3 is a flowchart illustrating the execution of a labeling subtask according to an embodiment of this application;
[0070] Figure 4 is a schematic diagram of an interface for configuring a labeling model provided in an embodiment of this application;
[0071] Figure 5 is a schematic diagram of a storage target task queue provided in an embodiment of this application;
[0072] Figure 6 is a flowchart illustrating a detection annotation subtask provided in an embodiment of this application;
[0073] Figure 7 is a schematic diagram of the architecture of an image annotation method provided in an embodiment of this application;
[0074] Figure 8 is a schematic diagram of the architecture of an image annotation system provided in an embodiment of this application;
[0075] Figure 9 is a schematic diagram of the architecture of another image annotation system provided in an embodiment of this application;
[0076] Figure 10 is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0077] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0078] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0079] As data annotation is applied in more and more scenarios, the annotation scenarios it faces are becoming increasingly complex. For example, in the field of autonomous driving, accurate image annotation can help vehicles identify road features, traffic lights, vehicles, pedestrians, and traffic signs, thereby making correct driving decisions. In the field of security monitoring, it can be used to annotate targets for inspection and behavior classification; in the field of medical imaging, it can be used to annotate features such as inspection areas, lesion areas, or tumors in images.
[0080] Figure 1 is a schematic diagram of an application scenario provided by an embodiment of this application. Referring to Figure 1, user equipment 102 and server 102 are shown. An image annotation system can be installed in user equipment 101. Users can generate image annotation requests through user equipment 101, and server 102 can receive image annotation requests and perform annotation processing on preset images according to the image annotation requests.
[0081] In related technologies, an image annotation request may include multiple sub-requests. After the annotation request is uploaded to the image annotation system, the image annotation system will determine the sub-task corresponding to each sub-request. Different sub-tasks are independent of each other. After the sub-task corresponding to each sub-request is completed, the next sub-task to be executed needs to be determined manually until all sub-tasks are completed. Then, the intermediate annotated images corresponding to each sub-task are selected manually and integrated to obtain the target annotated image.
[0082] The image annotation method provided in this application embodiment includes an image annotation request that may include the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request. Based on the configuration information corresponding to each annotation sub-request, annotation sub-tasks corresponding to each annotation sub-request can be generated, and each annotation sub-task is stored in a target task queue according to the execution order of the annotation sub-requests. Each annotation sub-task is executed according to its storage order in the target task queue to annotate a preset image, thereby obtaining an annotated image. This execution process can automatically execute annotation tasks through the target task queue to obtain an annotated image, eliminating the need for manual concatenation between sub-tasks and improving the efficiency of image annotation processing.
[0083] Figure 2 is a flowchart illustrating an image annotation method provided in an embodiment of this application. Referring to Figure 2, the method may include:
[0084] S201, Receive image annotation request.
[0085] Image annotation requests can be used to annotate preset images.
[0086] The preset image can be a user-specified image to be labeled. Users can upload the preset image along with the image labeling request, or specify the preset image from the preset storage device.
[0087] The preset storage device can store images for sharing. When image annotation is needed, the preset image can be directly obtained from the preset storage device, which can realize the flow and sharing of data.
[0088] In some possible embodiments, after a preset image has been annotated once, the annotated image can be stored in a preset storage device. When the annotated image needs to be annotated again in the future, the annotated image can be identified as the next preset image and annotated again.
[0089] In some possible embodiments, if the image annotation request and its corresponding preset image exist historically, the annotated image corresponding to the image annotation request is retrieved from a preset storage device. The preset storage device can be directly read and processed for subsequent repetitive and similar needs, reducing redundant annotation and avoiding waste of annotation resources.
[0090] An image annotation request can include the execution order of multiple annotation sub-requests and the configuration information for each annotation sub-request.
[0091] The image annotation system includes an image user interface, through which users can configure the configuration information corresponding to each annotation sub-request and determine the execution order of multiple annotation sub-requests.
[0092] After configuring each annotation sub-request, users can adjust the execution order of the annotation sub-requests by dragging and dropping components in the graphical user interface, thus determining the execution order of multiple annotation sub-requests.
[0093] In some possible embodiments, the execution order may include parallel execution and / or serial execution. For example, parallel execution may be used to indicate that at least two annotation sub-requests can simultaneously annotate a preset image, while serial execution may be used to indicate that the next annotation sub-request is executed only after one annotation sub-request has completed its annotation.
[0094] S202. Generate the annotation sub-task corresponding to each annotation sub-request based on the configuration information corresponding to each annotation sub-request.
[0095] Each annotation sub-request can annotate at least one annotation object. The configuration information corresponding to the annotation sub-request can include at least one annotation object and the annotation strategy corresponding to each annotation object. Annotation objects can be polylines, traffic lights, passable areas, and turning arrows, etc.
[0096] For any given labeled object, the labeling strategy can include a basic setting strategy and an attribute setting strategy. The basic setting strategy is used to indicate the storage fields of the category to which the labeled object belongs, and the attribute setting strategy is used to indicate the strategy information corresponding to each attribute in the labeled object.
[0097] The basic settings strategy may include at least one of the following: category name, category field, and subcategory type. The attribute settings strategy may include at least one of the following: attribute name, attribute field, attribute type, initial value, attribute configuration, and custom attribute configuration.
[0098] In some possible implementations, for any annotation sub-request, the annotation object corresponding to the annotation sub-request and the annotation strategy corresponding to the annotation object can be determined according to the configuration information; and the annotation sub-task corresponding to the annotation sub-request can be generated according to the annotation object and the annotation strategy corresponding to the annotation object.
[0099] A labeling subtask can be used to label at least one labeling object corresponding to a labeling subrequest.
[0100] For example, assuming there are 4 annotation sub-requests, namely annotation sub-requests 1-4, annotation sub-task 1 corresponding to annotation sub-request 1, annotation sub-task 2 corresponding to annotation sub-request 2, annotation sub-task 3 corresponding to annotation sub-request 3, and annotation sub-task 4 corresponding to annotation sub-request 4 can be generated.
[0101] S203. Store each annotation subtask in the target task queue according to the execution order corresponding to each annotation sub-request.
[0102] In some possible embodiments, at least one preset subtask and the execution order of the at least one preset subtask in the image annotation task are obtained; a subqueue corresponding to each annotation sub-request is determined according to the execution order of each annotation sub-request; the storage location of the subqueue in the target task queue is determined according to the execution order of at least one preset subtask in the image annotation task; and each annotation subtask is stored in the target task queue according to the storage location.
[0103] When annotating an image, in addition to the preset annotation subtasks, there are also fixed subtasks that are executed. The preset subtasks can be the subtasks in the image annotation task other than the various annotation subtasks.
[0104] For example, pre-defined subtasks may include point cloud pre-check subtasks, single-pass point cloud subtasks, dynamic keyframe subtasks, and multi-pass point cloud subtasks. Among them, the point cloud pre-check subtask can be used for data preprocessing and quality control to ensure the compliance of annotation input; the single-pass point cloud subtask can be used to process independent point cloud frames from a single acquisition; the dynamic keyframe subtask can be used for temporal annotation optimization of dynamic objects to reduce annotation redundancy through keyframes; and the multi-pass point cloud subtask is used for fusion and global annotation of multiple data acquisitions.
[0105] For example, suppose there are four annotation sub-requests, labeled sub-requests 1-4, where sub-request 1 corresponds to sub-task 1, sub-request 2 to sub-task 2, sub-request 3 to sub-task 3, and sub-request 4 to sub-task 4. Assuming the execution order is a serial execution of sub-requests 1, 2, 3, and 4, then the sub-queue is {sub-task 1, sub-task 2, sub-task 3, and sub-task 4}. Alternatively, suppose there are three preset sub-tasks, labeled sub-tasks 1-3, and at least one preset sub-task is executed in the image annotation task in the order of preset sub-task 1, 2, and 3. Based on the execution order of at least one preset sub-task in the image annotation task, it can be determined that the sub-queue is stored between preset sub-task 2 and preset sub-task 3 in the target task queue. Therefore, the target task queue is {preset sub-task 1, preset sub-task 2, annotation sub-task 1, annotation sub-task 2, annotation sub-task 3, annotation sub-task 4, preset sub-task 3}.
[0106] The execution order of at least one preset subtask in the image annotation task may include the execution sequence number corresponding to each preset subtask and the execution sequence number corresponding to the subqueue.
[0107] The storage location of the subqueue in the target task queue can be determined based on the execution sequence number corresponding to the subqueue.
[0108] In this application, necessary annotation subtasks in each image annotation task can be preset in advance by pre-setting subtasks. When a user's image annotation request is received, the sub-queue of the annotation subtask corresponding to the image annotation request can be stored in a suitable location in the image annotation task. The target task queue can be generated directly, without having to generate preset subtasks every time. This can improve the generation efficiency of the target task queue, thereby improving the annotation efficiency of the image annotation.
[0109] In some possible implementations, the target task queue can be stored as a queue template. When performing the same annotation again, the user can directly select the corresponding target task queue from the queue template, which can improve the efficiency of image annotation.
[0110] S204. Execute each annotation subtask according to the storage order of each annotation subtask in the target task queue to annotate the preset image and obtain an annotated image.
[0111] The target task queue may include multiple sub-queues, and each sub-queue includes at least one preset sub-task and each labeled sub-task.
[0112] The target task queue can be a serial task queue or a parallel task queue. In a serial task queue, subtasks can be executed sequentially from the head of the queue to the tail of the queue. In a parallel task queue, subtasks can be executed sequentially from the head of the queue to the tail of the queue. When executing at least two parallel labeled subtasks, two threads can be started to execute at least two labeled subtasks.
[0113] It can send labeled images to the user's device and store the labeled images to a preset storage device.
[0114] The image annotation method provided in this application can generate annotation sub-tasks corresponding to each annotation sub-request based on the configuration information corresponding to each annotation sub-request, and store each annotation sub-task in a target task queue according to the execution order of each annotation sub-request. Each annotation sub-task is executed according to its storage order in the target task queue to annotate a preset image and obtain an annotated image. Annotation tasks can be automatically executed through the target task queue to obtain an annotated image, eliminating the need for manual concatenation between sub-tasks and improving processing efficiency. Furthermore, preset sub-tasks can be automatically concatenated with each annotation sub-task to form a complete image annotation process, eliminating the need for manual intervention during image annotation and improving annotation accuracy.
[0115] Each annotation subtask can be executed to annotate a preset image and obtain an annotated image. For any annotation sub-request, the annotation sub-task can be executed through the annotation model corresponding to the annotation sub-task to achieve image annotation. The execution process of generating annotation sub-tasks provided in this application embodiment will be further explained below with reference to Figure 3.
[0116] Figure 3 is a flowchart illustrating the execution of a labeling subtask according to an embodiment of this application. Referring to Figure 3, the method may include:
[0117] S301. Obtain the annotation model corresponding to the annotation subtask from the annotation strategy corresponding to the annotation subtask.
[0118] An image annotation system can include multiple pre-defined annotation models. For example, a pre-defined model can be an image segmentation model (Segment Anything Model, SAM). SAM can achieve fully or semi-automatic image segmentation and annotation without domain-specific training through zero-shot learning and prompting engineering.
[0119] When configuring each annotation sub-request, users can configure the annotation model corresponding to each annotation sub-request. The annotation model can be selected from multiple preset annotation models, or a custom annotation model can be configured. For example, a custom annotation model can be the Open Neural Network Exchange (ONNX) model.
[0120] Figure 4 is a schematic diagram of an interface for configuring an annotation model according to an embodiment of this application. Referring to Figure 4, the left side is a navigation bar, which may include My Tasks, Annotation Platform Analysis, and Annotation Management. Annotation Management may include process nodes and process templates. After selecting a process node, annotation sub-requests can be configured. Configuration may include four settings: basic information, annotation procedures, advanced functions, and quality inspection processes. Annotation procedures allow configuration of the information corresponding to the annotation sub-request. The interface for advanced functions of a process node may include automation tool configuration and input / output format validation. Automation tool configuration includes multiple preset annotation models. Users can select the annotation model corresponding to the annotation sub-request from these preset models (preset annotation model 1, preset annotation model 2, and preset annotation model 3 in the figure). The input / output format validation bar includes an input format bar and an output format bar. After selecting to enter text in the input format bar and output format bar, users can customize the annotation model in the input format bar and output format bar instead of selecting the annotation model in the automation tool configuration bar.
[0121] In some possible implementations, users can configure multiple annotation models, which can assist each other in annotation, thereby improving annotation efficiency and accuracy.
[0122] S302. Execute the annotation sub-task through the annotation model corresponding to the annotation sub-task to perform annotation processing on the preset image and obtain the annotated image.
[0123] The image annotation method provided in this application can call the annotation model corresponding to the annotation subtask to annotate the image according to the configured annotation model, which can improve the efficiency of image annotation.
[0124] The image annotation system includes a task execution module, which contains an annotation queue. The task execution module can be used to execute multiple task queues in the annotation queue.
[0125] After generating the target task queue, the target task queue can be stored in the initial annotation queue. The following describes the execution process of storing the target task queue in this application with reference to Figure 5.
[0126] Figure 5 is a flowchart illustrating a method for storing a target task queue according to an embodiment of this application. Referring to Figure 5, the method may include:
[0127] S501. Obtain the target priority corresponding to the target task queue.
[0128] An image processing request can include a target priority, which is a priority configured by the user for the image processing request. For example, the priority can be selected from low, medium, and high priority. The target priority is related to the storage position of the target task queue within the annotation queue. For two task queues to be executed, if their priorities are different, the task queue with the higher priority will be stored earlier in the list.
[0129] Users can be prompted to choose between low and medium priority, reserving the high priority for use in subsequent urgent adjustment tasks.
[0130] S502. If it is determined that the storage of the target task queue in the initial labeling queue does not meet the target priority requirements, then update the storage location of the target task queue in the initial labeling queue to obtain the target labeling queue.
[0131] The target labeling queue includes multiple queues of tasks to be executed, and the target task queue is one of these queues.
[0132] For example, the initial annotation queue includes task queue 1 and task queue 2, where task queue 1 is of medium priority and task queue 2 is of low priority. Task queue 1 is at the front of the queue, and task queue 2 is at the back. Assuming the target task queue is task queue 3, after storing task queue 3 into the initial annotation queue, task queue 1 is at the front, task queue 2 is in the middle, and task queue 3 is at the back. If task queue 3 is of low priority, then the target priority requirement is met, and no update to the annotation queue is needed. If task queue 3 is of medium priority, then the target priority requirement is not met, and task queue 3 is moved to the middle of the queue, and task queue 2 is moved to the back, resulting in the target annotation queue.
[0133] When executing each task queue in the labeled queue sequentially, there may be task queues that need to be modified. After the user reconfigures the task queue, a task change request corresponding to the target task queue can be received, and the target task queue can be modified accordingly.
[0134] Task change requests can be submitted by the user corresponding to the target task queue or by the administrator. Task change requests can include at least one of the following: priority change and subtask labeling change.
[0135] In some possible implementations, if the target task queue consumes a significant amount of memory resources while the resources available for the entire image annotation system are limited, the efficiency of task annotation in the target task queue will be low. In such cases, an alarm can be issued for the target task queue, indicating that current resources are limited and the target task queue is consuming excessive memory. Based on the alarm, administrators can determine a task change request and modify the priority of the target task queue.
[0136] If the task change request requests a priority change for the target task queue, the task change request includes the change priority corresponding to the target task queue; based on the change priority, the storage location of the target task queue is updated in the target annotation queue.
[0137] If the task change request requests a change to a labeled subtask in the target task queue, the task change request may include the change information corresponding to the labeled subtask; based on the change information, the labeled subtask in the target task queue is updated.
[0138] It's worth noting that after changes are made to the target task queue, the current version of the target task queue can be modified on the visual interface. For example, the previous version was V1, and the current version is V2. Multiple version information for each target queue can be recorded.
[0139] The image annotation method provided in this application can automatically allocate and optimize annotation tasks and dynamically adjust the task order when the number of annotation tasks to be processed is very large, thereby improving the processing efficiency of the tasks. Furthermore, the priority of annotation tasks and annotation sub-tasks can be changed according to task change requests, achieving flexibility in image annotation.
[0140] During the execution of the target task queue, the execution status of each labeled subtask can be detected. The following is a detailed explanation with reference to Figure 6.
[0141] Figure 6 is a flowchart illustrating a detection annotation subtask provided in an embodiment of this application. Referring to Figure 6, the method may include:
[0142] S601. Detect the verification identifier of each target subtask in the target task queue.
[0143] The verification identifiers for the target subtask can be either a verification exception identifier or a verification pass identifier.
[0144] In some possible embodiments, after the target subtask is completed, the intermediate image labeled by the target subtask can be detected by the detection model to detect whether the intermediate image meets the labeling requirements. If it meets the labeling requirements, the verification flag is updated to the verification pass flag; if it does not meet the labeling requirements, the verification flag is updated to the verification failure flag.
[0145] The detection model can be one of several preset models or a custom model.
[0146] In some possible implementations, after an administrator discovers an anomaly in the target subtask, they can change the verification identifier corresponding to the target subtask.
[0147] S602. If the verification identifier of the target subtask is a verification exception identifier, the target subtask is identified as an exception subtask.
[0148] S603. Handle abnormal subtasks, including rolling back execution or skipping execution.
[0149] If the abnormal subtask is in an unexecuted state, it can be skipped and the next marked subtask can be executed.
[0150] If the abnormal subtask is already executed, it can be rolled back to re-execute the abnormal subtask.
[0151] The image annotation method provided in this application embodiment can detect the verification identifier of each subtask. If there is an abnormal subtask, it can be skipped or rolled back in a timely manner, which can improve the stability of the annotation process.
[0152] Figure 7 is a schematic diagram of the architecture of an image annotation method provided in an embodiment of this application. Referring to Figure 7, when there are image annotation requirements on the user end (requirements A, B, and N shown in Figure 7), the user can configure each requirement on the user terminal's visual interface to generate image annotation requests corresponding to each requirement (for example, image annotation request A corresponding to requirement A, image annotation request B corresponding to requirement B, and image annotation request N corresponding to requirement N in Figure 7). In the image annotation system, image annotation requests can be received, and images can be annotated according to the image annotation requests to obtain an annotated image (for example, an annotated image A corresponding to requirement A, an annotated image B corresponding to requirement B, and an annotated image N corresponding to requirement N in Figure 7). Finally, the annotated image is stored in a preset storage device.
[0153] The user interface allows for workflow management of sub-requests within an image annotation request, including workflow control, execution order calculation, annotation model settings, and pre-annotation settings. When users input custom configuration information, the interface also allows for matching the user-input configuration with the image annotation system. Furthermore, users can dynamically update the task queue corresponding to the image annotation request through task change requests.
[0154] Users can automatically annotate images by configuring image annotation requests through a visual interface. At the system's underlying management level, image annotation methods enable automatic concatenation of subtasks, subtask rollback / skipping, templated annotation tasks, and automatic execution of subtasks. Furthermore, the system can provide preset annotation models, allowing each subtask to run independently and automatically merge annotation results, achieving fully automated annotation.
[0155] Figure 8 is a schematic diagram of the architecture of an image annotation system provided in an embodiment of this application. Referring to Figure 8, the image annotation system 800 may include a request receiving module 801, a subtask generation module 802, a task queue generation module 803, and an annotation module 804, wherein...
[0156] The request receiving module 801 is used to receive image annotation requests;
[0157] The subtask generation module 802 is used to generate annotation subtasks corresponding to each annotation sub-request based on the configuration information corresponding to each annotation sub-request.
[0158] The task queue generation module 803 is used to store each annotation sub-task into the target task queue according to the execution order corresponding to each annotation sub-request;
[0159] The annotation module 804 is used to execute each annotation subtask according to the storage order of each annotation subtask in the target task queue, so as to perform annotation processing on the preset image and obtain an annotated image.
[0160] In some possible embodiments, the task queue generation module 803 is also used for:
[0161] Obtain at least one preset subtask and the execution order of at least one preset subtask in the image annotation task. The preset subtask is a subtask in the image annotation task other than each annotation subtask.
[0162] Based on the execution order of each annotation sub-request, determine the sub-queue corresponding to each annotation sub-task;
[0163] The storage location of the sub-queue in the target task queue is determined based on the execution order of at least one preset sub-task in the image annotation task;
[0164] Store each labeled subtask into the target task queue according to its storage location.
[0165] In some possible embodiments, the subtask generation module 802 is specifically used for:
[0166] For any given annotation sub-request, determine the annotation object corresponding to the annotation sub-request and the annotation strategy corresponding to the annotation object based on the configuration information;
[0167] For any annotation sub-request, generate the corresponding annotation sub-task based on the annotation object and the annotation strategy corresponding to the annotation object.
[0168] In some possible embodiments, the annotation module 804 is also used for:
[0169] Obtain the annotation model corresponding to the annotation subtask from the annotation strategy corresponding to the annotation subtask;
[0170] The annotation subtask is executed by the annotation model corresponding to the annotation subtask to annotate the preset image and obtain the annotated image.
[0171] Figure 9 is a schematic diagram of the architecture of another image annotation system provided in an embodiment of this application. Referring to Figure 9, the image annotation system 800 also includes a storage module 805, a detection module 806, and a modification module 807.
[0172] In some possible embodiments, storage module 805 is used to store the target task queue into the initial labeling queue, and storage module 804 has functions for:
[0173] Obtain the target priority corresponding to the target task queue;
[0174] If it is determined that the storage of the target task queue in the initial annotation queue does not meet the target priority requirements, then the storage location of the target task queue in the initial annotation queue is updated to obtain the target annotation queue. The target annotation queue includes multiple task queues to be executed, and the target task queue is included among the multiple task queues to be executed.
[0175] In some possible embodiments, the detection module 806 is used for:
[0176] The verification identifiers of each target subtask in the target task queue are checked;
[0177] If the verification identifier of a target subtask is a verification exception identifier, the target subtask will be identified as an exception subtask.
[0178] Perform exception handling for abnormal subtasks, including rolling back execution or skipping execution.
[0179] In some possible embodiments, the change module 807 is used for:
[0180] Receive task change requests corresponding to the target task queue;
[0181] If the task change request requests a priority change for the target task queue, the task change request includes the change priority corresponding to the target task queue; based on the change priority, the storage location of the target task queue is updated in the target annotation queue;
[0182] If the task change request requests a change to a labeled subtask in the target task queue, the task change request includes the change information corresponding to the labeled subtask; based on the change information, the labeled subtask in the target task queue is updated.
[0183] The image annotation system provided in this application embodiment can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar, and will not be repeated here.
[0184] Figure 10 is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Referring to Figure 10, the electronic device 1000 may include a processor 1001 and a memory 1002. Exemplarily, the processor 1001 and the memory 1002 are interconnected via a bus 1003.
[0185] Memory 1002 stores computer-executed instructions;
[0186] The processor 1001 executes computer execution instructions stored in the memory 1002, causing the processor 1001 to perform the image annotation method as shown in the above method embodiment.
[0187] Accordingly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the image annotation method of the above-described method embodiments.
[0188] Accordingly, embodiments of this application may also provide a computer program product, including a computer program, which, when executed by a processor, can implement the image annotation method shown in the above method embodiments.
[0189] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0190] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more flowchart illustrations and / or one or more block diagrams.
[0191] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.
[0192] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.
[0193] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0194] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0195] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, 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 disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0196] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0197] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. An image annotation method, characterized in that, include: Receive an image annotation request, the image annotation request is used to annotate a preset image, the image annotation request includes the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request; Based on the configuration information corresponding to each annotation sub-request, generate annotation sub-tasks corresponding to each annotation sub-request, and store each annotation sub-task in the target task queue according to the execution order of each annotation sub-request; Each annotation subtask is executed in the storage order of the target task queue to annotate the preset image and obtain the annotated image.
2. The method according to claim 1, characterized in that, Before storing each annotation subtask into the target task queue according to the execution order corresponding to each annotation sub-request, the method further includes: Obtain at least one preset subtask and the execution order of the at least one preset subtask in the image annotation task. The preset subtask is a subtask in the image annotation task other than each annotation subtask. Based on the execution order of each annotation sub-request, determine the sub-queue corresponding to each annotation sub-task; The storage location of the sub-queue in the target task queue is determined according to the execution order of at least one of the preset sub-tasks in the image annotation task; Each labeled subtask is stored in the target task queue according to the specified storage location.
3. The method according to claim 2, characterized in that, The execution order of the at least one preset subtask in the image annotation task includes the execution sequence number corresponding to each preset subtask and the execution sequence number corresponding to the subqueue. Determining the storage location of the subqueue in the target task queue based on the execution order of the at least one preset subtask in the image annotation task includes: The storage location of the sub-queue in the target task queue is determined based on the execution sequence number corresponding to the sub-queue.
4. The method according to any one of claims 1-3, characterized in that, The step of generating a labeling sub-task corresponding to each labeling sub-request based on the configuration information corresponding to each labeling sub-request includes: For any given annotation sub-request, the annotation object corresponding to the annotation sub-request and the annotation strategy corresponding to the annotation object are determined based on the configuration information. For any given annotation sub-request, an annotation sub-task corresponding to the annotation sub-request is generated based on the annotation object and the annotation strategy corresponding to the annotation object.
5. The method according to any one of claims 1-4, characterized in that, For any given annotation subtask, each annotation subtask is executed to annotate the preset image, resulting in the annotated image, including: Obtain the annotation model corresponding to the annotation subtask from the annotation strategy corresponding to the annotation subtask; The annotation subtask is executed by the annotation model corresponding to the annotation subtask to annotate the preset image and obtain the annotated image.
6. The method according to any one of claims 1-5, characterized in that, The target task queue is stored in the initial labeling queue; Before executing each annotation subtask in the target task queue according to the storage order of each annotation subtask to annotate the preset image and obtain the annotated image, the method further includes: Obtain the target priority corresponding to the target task queue; If it is determined that the storage location of the target task queue in the initial annotation queue does not meet the target priority requirement, then the storage location of the target task queue in the initial annotation queue is updated to obtain the target annotation queue. The target annotation queue includes multiple task queues to be executed, and the target task queue is included among the multiple task queues to be executed.
7. The method according to any one of claims 1-6, characterized in that, The method further includes: The verification identifiers of each target subtask in the target task queue are detected; If the verification identifier of a target subtask is a verification anomaly identifier, the target subtask is identified as an anomaly subtask; The abnormal subtask is handled in an exception manner, which includes rolling back execution or skipping execution.
8. The method according to claim 7, characterized in that, The detection of the verification identifier of each target subtask in the target task queue includes: The detection model is used to detect the intermediate images after the target subtask is labeled, and to detect whether the intermediate images meet the labeling requirements. If the labeling requirements are met, the verification identifier will be updated to a verification passed identifier; If the labeling requirements are not met, the verification label will be updated to a verification anomaly label.
9. The method according to claim 7, characterized in that, The exception handling for the abnormal subtask includes: If the abnormal subtask is in an unexecuted state, then the abnormal subtask is skipped and the next labeled subtask is executed. If the abnormal subtask is already executed, then the abnormal subtask is rolled back to be re-executed.
10. The method according to claim 6, characterized in that, The method further includes: Receive the task change request corresponding to the target task queue; If the task change request is a request to change the priority of the target task queue, the task change request includes the change priority corresponding to the target task queue; according to the change priority, the storage location of the target task queue is updated in the target label queue; If the task change request is a request to change the labeled subtasks of the target task queue, the task change request includes change information corresponding to the labeled subtasks; and the labeled subtasks of the target task queue are updated according to the change information.
11. The method according to claim 1, characterized in that, The receiving of image annotation requests includes: In response to the configuration of image annotation requirements, the image annotation request is generated; wherein the configuration information of the image annotation requirements includes: at least one annotation object and an annotation strategy corresponding to each annotation object.
12. The method according to claim 11, characterized in that, The configuration of image annotation requirements includes at least one of the following: The process management is performed on the annotation sub-requests corresponding to the image annotation requirements. The execution order of each labeled sub-request is calculated and arranged; Configure the annotation model for each annotation sub-request; The input configuration is pre-annotated, which is used to match the configuration information input by the user with the image annotation system.
13. An image annotation system, characterized in that, The image annotation system includes a request receiving module, a subtask generation module, a task queue generation module, and an annotation module, wherein... The request receiving module is used to receive image annotation requests, which are used to annotate a preset image. The image annotation request includes the execution order of multiple annotation sub-requests and the configuration information corresponding to each annotation sub-request. The subtask generation module is used to generate annotation subtasks corresponding to each annotation sub-request based on the configuration information corresponding to each annotation sub-request. The task queue generation module is used to store each annotation subtask into the target task queue according to the execution order corresponding to each annotation sub-request; The annotation module is used to execute each annotation subtask according to the storage order of each annotation subtask in the target task queue, so as to perform annotation processing on the preset image and obtain the annotated image.
14. The image annotation system according to claim 13, characterized in that, The system also includes a storage module; The storage module is used to obtain the target priority corresponding to the target task queue; If it is determined that the storage of the target task queue in the initial annotation queue does not meet the target priority requirements, then the storage location of the target task queue is updated in the initial annotation queue to obtain the target annotation queue. The target annotation queue includes multiple task queues to be executed, and the target task queue is included among the multiple task queues to be executed.
15. The image annotation system according to claim 13 or 14, characterized in that, The system also includes a detection module for: The verification identifiers of each target subtask in the target task queue are checked; If the verification identifier of a target subtask is a verification exception identifier, the target subtask will be identified as an exception subtask. Perform exception handling for abnormal subtasks, including rolling back execution or skipping execution.
16. The image annotation system according to any one of claims 13-15, characterized in that, The system also includes a change module, used for: Receive task change requests corresponding to the target task queue; If the task change request requests a priority change for the target task queue, the task change request includes the priority change corresponding to the target task queue; Based on the change priority, update the storage location of the target task queue in the target labeling queue; If the task change request requests a change to a labeled subtask in the target task queue, the task change request includes the change information corresponding to the labeled subtask. Based on the change information, update the labeled subtasks in the target task queue.
17. An electronic device, characterized in that, include: Memory and processor The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the image annotation method as described in any one of claims 1 to 12.
18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the image annotation method according to any one of claims 1 to 12.
19. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the image annotation method according to any one of claims 1 to 12.
