CN113254221A - Task execution system and method for defect labeling - Google Patents
Task execution system and method for defect labeling Download PDFInfo
- Publication number
- CN113254221A CN113254221A CN202110778856.1A CN202110778856A CN113254221A CN 113254221 A CN113254221 A CN 113254221A CN 202110778856 A CN202110778856 A CN 202110778856A CN 113254221 A CN113254221 A CN 113254221A
- Authority
- CN
- China
- Prior art keywords
- task
- data set
- task execution
- information
- unmarked
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a task execution system and method for defect marking, wherein a task sending end is used for determining marking information and authority information of a marked task, so that the task execution end can only check unmarked data set information corresponding to the distributed task, and after the task execution end receives a task distribution instruction, only can check the unmarked data set information corresponding to the task, check the unmarked data set of the unmarked data set information, and mark the unmarked data set, thereby improving the safety of the unmarked data set information and the marked task.
Description
Technical Field
The invention relates to the technical field of task authority management, in particular to a task execution system and method for defect marking.
Background
With the development of Artificial Intelligence (AI) technology, various deep learning platforms for providing a service of training and reasoning a deep learning model have emerged. The deep learning of needing to train the deep learning model and reasoning service provides a large amount of label data, and the current can not provide a large amount of label data fast, if treat the label data execution label task, suffer data leakage easily, can't guarantee to treat the security of label data information and label task.
Disclosure of Invention
The invention mainly aims to provide a task execution system and a task execution method for defect marking, and aims to solve the technical problems that the existing task for executing the marking of data to be marked is easy to leak data, and the safety of the information of the data to be marked and the marking task cannot be ensured.
In a first aspect, the present invention provides a task execution system for defect labeling, the system comprising the following steps:
the task sending end and the task executing end;
the task sending end determines marking information, wherein the marking information at least comprises unmarked data set information required by executing a task and is also used for determining authority information, so that the task executing end can only check the unmarked data set information corresponding to the distributed task and then send a task distribution instruction;
the task execution end receives a task allocation instruction, checks an unmarked data set according to unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
In a second aspect, the present invention further provides a task execution system method for defect labeling, the method comprising the following steps:
determining annotation information, wherein the annotation information at least comprises information of an unmarked data set required by executing a task;
determining authority information to enable the task execution end to only view the unmarked data set information corresponding to the distributed tasks;
and sending a task allocation instruction, so that the task execution end checks the unmarked data set according to the unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
Optionally, the task execution end includes a plurality of task execution ends, including:
sending corresponding tasks to a plurality of task execution ends, and distributing unmarked data set information required by the task execution to the plurality of task execution ends;
and enabling a plurality of task execution ends to cooperatively complete the tasks according to the distributed information of the unmarked data sets.
Optionally, the number of the task execution ends is multiple, the task execution condition of each task execution end is monitored in real time, and the task amount of each task execution end is adjusted according to the task execution condition.
Optionally, the task execution end includes a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: and if the fact that the first task execution end does not finish the labeling task of the unmarked data set information distributed when the first task execution end is offline is determined, task recovery is carried out on the labeling task of the unmarked data set information, and the unmarked data set information is distributed to other online task execution ends as a new task, wherein the unmarked data set information comprises the recovered unmarked data set information.
Optionally, the task execution end includes a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: if the first task execution end is determined to finish the labeling task of the distributed unlabeled data set information, the unlabeled data set information which is not labeled by the other task execution ends is recycled, and the unlabeled data set information is distributed to the first task execution end as a new task, wherein the unlabeled data set information comprises the recycled unlabeled data set information.
Optionally, after the task allocation instruction is sent, the method further includes:
and displaying at least one of undistributed data set information, data set information of which the task execution end finishes labeling and data set information of which the task execution end does not finish labeling.
Optionally, the unlabeled data set information at least includes a storage path of the unlabeled data set.
In a third aspect, the present invention further provides a task execution system method for defect labeling, including the following steps:
receiving a task allocation instruction sent by a task sending end, and acquiring unmarked data set information carried by the allocation instruction;
and checking the unmarked data set based on the unmarked data set information, and marking the unmarked data set.
Optionally, after the annotating the unlabeled data set, the method further includes:
and feeding back the off-line information so that the task management end can recycle the data set information which is not marked in the task.
The invention provides a task execution system and a method for defect marking, wherein the marking information and the authority information of a marked task are determined by a task sending end, so that the task execution end can only check the unmarked data set information corresponding to the distributed task, and after the task execution end receives a task distribution instruction, only the unmarked data set information corresponding to the task can be checked, the unmarked data set of the unmarked data set information can be checked, and the unmarked data set can be marked, thereby improving the safety of the unmarked data set information and the marked task; the invention sends the information of the unmarked data set, the unmarked data set is not directly sent, and the task execution end only checks the unmarked data set according to the unmarked data set information carried by the task allocation instruction so as to carry out marking, thereby reducing the occupied bandwidth; according to the invention, the tasks are automatically distributed to the multiple execution terminals, so that the multiple terminals can execute the tasks at the same time, and the task completion efficiency is improved; the invention monitors the task execution condition of each task execution end in real time and adjusts the task quantity of each task execution end according to the task execution condition, thereby balancing the task quantity of each execution terminal and being beneficial to improving the task completion efficiency.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart of a task execution system for defect labeling according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a system scenario provided in the present embodiment;
fig. 3 is a flowchart illustrating another task execution method for defect labeling according to an embodiment of the present invention;
fig. 4 is a flowchart illustrating another task execution method for defect labeling according to an embodiment of the present invention;
fig. 5 is a block diagram schematically illustrating a structure of a computer device according to an embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The flow diagrams depicted in the figures are merely illustrative and do not necessarily include all of the elements and operations/steps, nor do they necessarily have to be performed in the order depicted. For example, some operations/steps may be decomposed, combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
The embodiment of the invention provides a task execution system and method for defect marking. The task execution method for defect labeling can be applied to computer equipment, and the computer equipment can be electronic equipment such as a notebook computer and a desktop computer.
Some embodiments of the invention are described in detail below with reference to the accompanying drawings. The embodiments described below and the features of the embodiments can be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a flowchart illustrating a task execution system for defect labeling according to an embodiment of the present invention.
As shown in fig. 1, the method includes steps S10 through S12.
Step S10, the task sending end and the task executing end.
Exemplarily, the task sending end is a terminal operated by a manager, the task executing end is a terminal operated by an operator, and the task sending end and the task executing end can be electronic devices such as a desktop computer and a notebook computer.
Step S11, the task sending end determines the annotation information, where the annotation information at least includes the unmarked data set information required for executing the task, and is also used to determine the authority information, so that the task executing end can only view the unmarked data set information corresponding to the assigned task, and then sends a task assignment instruction.
Exemplarily, the task sending end determines the labeling information, where the labeling information at least includes information of an unlabeled data set required for executing the task, and further includes information of a defect label required for data labeling, that is, what kind of defect is to be labeled. For example, as shown in fig. 2, the unlabeled data set 1 and/or the unlabeled data set 2 are determined from the unlabeled data sets in the server, such as the unlabeled data set 1 and the unlabeled data set 2, as data sets required by the task, and information of the unlabeled data set 1 is determined, such as a storage path address of the unlabeled data set 1, a name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, and names of the pictures, etc. are determined; and/or determining the information of the unlabeled data set 2, such as looking at the storage path address of the unlabeled data set 2, the name of the unlabeled data set 2, the number of pictures in the unlabeled data set 2, and the name information of each picture. The label information required by data labeling can also be determined, for example, the asd label information is determined to be the defect label information required to be labeled in the unlabeled data set 1 and/or the unlabeled data set 2; and/or determining the white label information as the defect label information required for labeling the unlabeled data set 1 and/or the unlabeled data set 2.
Taking the determined unmarked data set 1 and/or the unmarked data set 2 as an unmarked task to be executed, and setting the authority of the marked task, so that the task execution end marked with the unmarked data set 1 can only view the unmarked data set 1 and the related information thereof, including a storage path address, the name of the unmarked data set 1, the number of pictures in the unmarked data set 1, the name of each picture, and asd label information and white label information which need to be marked in the unmarked data set 1; and/or checking the unlabeled data set 2 and related information thereof, including a storage path address, a name of the unlabeled data set 2, the number of pictures in the unlabeled data set 2, a name of each picture, and asd label information and white label information to be labeled of the unlabeled data set 2. As shown in fig. 2, the task sending end sends a task allocation instruction to the task executing end.
And step S12, the task execution end receives a task allocation instruction, checks an unmarked data set according to unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
Exemplarily, for example, as shown in fig. 2, the task execution end receives a task allocation instruction sent by the task sending end, and views the allocated unlabeled data set and related information according to the task allocation instruction, for example, the task execution end of the unlabeled data set 1 information can only view the unlabeled data set 1 and related information thereof, including a storage path address of the unlabeled data set 1, a name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, names of the pictures, and asd tag information and while tag information required for labeling the unlabeled data set 1; and/or checking the unlabeled data set 2 and related information thereof, including the storage path address of the unlabeled data set 2, the name of the unlabeled data set 2, the number of pictures in the unlabeled data set 2, the name of each picture, and asd label information and white label information which need to be labeled by the unlabeled data set 2.
For example, the picture to be labeled is checked through the storage path address of the unlabeled data set 1 and the name of the unlabeled data set 1, the picture to be labeled in the unlabeled data set 1 is labeled through the asd label information and the white label information, and/or the picture to be labeled is checked through the storage path address of the unlabeled data set 2 and the name of the unlabeled data set 2, and the picture to be labeled in the unlabeled data set 2 is labeled through the asd label information and the white label information.
In the embodiment, the labeling information and the authority information of the labeling task are determined by the task sending end, so that the task execution end can only check the unlabeled data set information corresponding to the distributed task, and after the task execution end receives the task distribution instruction, only the unlabeled data set information and the unlabeled data set corresponding to the task can be checked, and the unlabeled data set is labeled, so that the safety of the unlabeled data set information and the labeling task is improved.
Referring to fig. 3, fig. 3 is a flowchart illustrating a task execution method for defect labeling according to an embodiment of the present invention.
As shown in fig. 3, this includes steps S20 to S22.
And step S20, determining annotation information, wherein the annotation information at least comprises information of the unmarked data set required for executing the task.
Exemplarily, the task sending end determines the labeling information, where the labeling information at least includes information of an unlabeled data set required for executing the task, and further includes information of a defect label required for data labeling, that is, what kind of defect is to be labeled. For example, as shown in fig. 2, the unlabeled data set 1 and/or the unlabeled data set 2 are determined from the unlabeled data sets in the server, such as the unlabeled data set 1, the unlabeled data set 2, and the like, as data sets required by the task, and the unlabeled data set 1 information, such as the storage path address of the unlabeled data set 1, the name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, the name of each picture, and the like, is determined; and/or determining the information of the unlabeled data set 2, such as the storage path address of the unlabeled data set 2, the name of the unlabeled data set 2, the number of pictures in the unlabeled data set 2, and the name information of each picture. The label information required by data labeling can also be determined, for example, the asd label information is determined to be the defect label information required to be labeled in the unlabeled data set 1 and/or the unlabeled data set 2; and/or determining that the white tag information is the defect tag information which needs to be labeled for the unlabeled data set 1 and/or the unlabeled data set 2.
And step S21, determining authority information, so that the task execution end can only view the unmarked data set information corresponding to the distributed tasks.
Exemplarily, setting the authority of the labeling task, so that the task execution end labeled with the unlabeled data set 1 can only view the unlabeled data set 1 and related information thereof, including a storage path address, a name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, names of the pictures, and asd tag information and white tag information to be labeled of the unlabeled data set 1; and/or checking the unlabeled data set 2 and related information thereof, including a storage path address, a name of the unlabeled data set 2, the number of pictures in the unlabeled data set 2, a name of each picture, and asd label information and white label information to be labeled of the unlabeled data set 2. As shown in fig. 2, the task sending end sends a task allocation instruction to the task executing end.
And step S22, sending a task allocation instruction, enabling the task execution end to check the unmarked data set according to the unmarked data set information carried by the task allocation instruction, and marking the unmarked data set.
Exemplarily, as shown in fig. 2, the task sending end sends a task allocation instruction to the task executing end. And displaying at least one or more of undistributed data set names, undistributed picture numbers, annotated data set names, annotated picture numbers, unlabeled data set names, unlabeled picture numbers, annotated task execution end information of annotated tasks and annotated task execution end information of the unlabeled tasks on a task allocation interface.
The task execution end receives a task allocation instruction sent by the task sending end, and checks the label information in the task allocation instruction, if the label information is the information of the unlabeled data set 1, the task execution end can only check the picture information of the unlabeled data set 1, the storage path address of the unlabeled data set 1, the name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, the name of each picture, and asd label information and white label information which need to be labeled in the unlabeled data set 1; and/or the marked information is the information of the unmarked data set 2, and the task execution end can only check the picture information of the unmarked data set 2, the storage path address of the unmarked data set 2, the name of the unmarked data set 2, the number of pictures in the unmarked data set 2, the name of each picture, and asd tag information and white tag information which need to be marked by the unmarked data set 2.
For example, the task execution end checks the pictures to be labeled through the storage path address of the unlabeled data set 1 and the name of the unlabeled data set 1, labels the pictures to be labeled in the unlabeled data set 1 through the asd label information and the white label information, and/or checks the pictures to be labeled through the storage path address of the unlabeled data set 2 and the name of the unlabeled data set 2, and labels the pictures to be labeled in the unlabeled data set 2 through the asd label information and the white label information. It should be noted here that the task execution end may directly label the checked unlabeled data set without downloading the allocated unlabeled data set.
In one embodiment, the task execution end includes a plurality of ends, including: sending corresponding tasks to a plurality of task execution ends, and distributing unmarked data set information required by the task execution to the plurality of task execution ends; and enabling a plurality of task execution ends to cooperatively complete the tasks according to the distributed information of the unmarked data sets.
Illustratively, for example, the task execution terminals include at least two, such as a first task execution terminal and a second task execution terminal. When the task sending end looks at the number of pictures in the unlabelled data set 1, the picture information in the unlabelled data set 1 for executing the labeled task is distributed to the first task execution end and the second task execution end according to the number of pictures, for example, when the number of pictures in the unlabelled data set 1 is 100, 50 pieces of picture information are distributed to the first task execution end, and the other 50 pieces of picture information are distributed to the second task execution end. Or distributing 25 pieces of picture information to the first execution task end, and distributing 25 pieces of picture information to the second execution task end. The picture information at least comprises a path address, and can also comprise a name of the unlabeled data set 1, a name of a picture, and asd label information and white label information labeled on the picture. And distributing the tasks to a plurality of execution ends to improve the task completion efficiency.
The first task execution end and the second task execution end receive task allocation instructions belonging to the first task execution end and the second task execution end respectively, view the allocated picture information and label information used for labeling pictures, such as an asd label and a white label, view the pictures needing to be labeled through the picture information in a collaborative labeling mode, and use the asd label and the white label to perform defect labeling on the pictures needing to be labeled, so that the allocated tasks are completed.
In an embodiment, the task execution ends include a plurality of task execution ends, the task execution condition of each task execution end is monitored in real time, and the task amount of each task execution end is adjusted according to the task execution condition.
Exemplarily, the execution condition of each task execution end executing the corresponding task is detected in real time, for example, the image information with the label completed and/or the image information with the label not completed is detected, the image information at least includes the path address, and may also include the name of the data set 1 that is not labeled, the name of the image, and asd label information and white label information labeled to the image. If the task volume marked with the 50 pieces of picture information is sent to the first task execution end and the second task execution end respectively, if the first task execution end is detected to be marked with 30 pieces of pictures and the second task execution end is detected to be marked with 10 pieces of pictures, 20 pieces of picture information which are not marked in the second task execution end are generated into a new marked task and are distributed to the first task execution end. Or sending the task amount of 25 pieces of picture information to the first task execution end and the second task execution end respectively, wherein the task sending end still retains 50 pieces of picture information to be labeled, and if the first task execution end is detected to label 10 pieces of pictures and the second task execution end is detected to label 5 pieces of pictures, the remaining 50 pieces of picture information to be labeled are continuously generated into a new task and are distributed to the first task execution end and the second task execution end, so that the first task execution end receives the 10 pieces of picture information to be labeled in the new task, and the second task execution end receives the 5 pieces of picture information to be labeled in the new task.
In one embodiment, the task execution end comprises a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: and if the fact that the first task execution end does not finish the labeling task of the unmarked data set information distributed when the first task execution end is offline is determined, task recovery is carried out on the labeling task of the unmarked data set information, and the unmarked data set information is distributed to other online task execution ends as a new task, wherein the unmarked data set information comprises the recovered unmarked data set information.
Exemplarily, the first executive task end is detected to be off-line, the current execution annotation task condition of the first executive task end is checked, if the first task execution end does not finish the labeling task of the distributed unlabelled data set information, the picture information to be labeled in the unlabelled data set information which does not finish the labeling task is recovered, the unlabelled data set information is generated into a new labeling task to be distributed to other online task execution ends, for example, 20 pieces of unlabelled picture information are collected in the unlabelled data set in the first task execution end, the 20 pieces of unlabelled picture information are collected into the unassigned unlabelled data set information, generating a new task by using the undistributed unlabelled data set information comprising the 20 pieces of image information which are not labeled, distributing the new task to the second online executing task end, or generating a new task by using 20 pieces of image information without labels and distributing the new task to the second online executing task end. And monitoring the task execution condition of each task execution end in real time, and adjusting the task quantity of each task execution end according to the task execution condition, so that the task quantity of each execution terminal is balanced, and the task completion efficiency is improved.
In one embodiment, the task execution end comprises a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: if the first task execution end is determined to finish the labeling task of the distributed unlabeled data set information, the unlabeled data set information which is not labeled by the other task execution ends is recycled, and the unlabeled data set information is distributed to the first task execution end as a new task, wherein the unlabeled data set information comprises the recycled unlabeled data set information.
Exemplarily, it is detected that the first task execution end completes the labeling task of the distributed unlabeled data set information, the picture information to be labeled in the unlabeled data set information of which the other online task execution ends do not complete the labeling task is recovered, the unlabeled data set information is generated into a new labeling task, and the new labeling task is distributed to the first task execution end, wherein the recovering comprises recovering quantitative unlabeled picture information, or recovering all unlabeled picture information. For example, the first task execution end completes the distributed labeling task, at this time, 20 pieces of unlabelled picture information are included in the unlabelled data set information in the second task execution end, the 20 pieces of unlabelled picture information in the second task execution end are recycled to the unassigned unlabelled data set information, a new task is generated and distributed to the online first execution task end by the unassigned unlabelled data set information including the 20 pieces of unlabelled picture information, or the 20 pieces of unlabelled picture information generate a new task and distribute the new task to the online first execution task end.
Or, the 10 pieces of image information which are not marked in the second task execution end are recycled to the unallocated unmarked data set information, and the unallocated unmarked data set information including the 10 pieces of image information which are not marked generates a new task to be allocated to the online first task execution end, or the 10 pieces of image information which are not marked generates the new task to be allocated to the online first task execution end. It should be emphasized here that it is only the picture related information that is recycled, and no picture is included, for example, 10 pieces of picture information that are not labeled in the second task execution end are recycled, including the storage path and the picture label of the 10 pieces of picture, and then the above information is generated into a new task to be sent to the online first task execution end. And monitoring the task execution condition of each task execution end in real time, and adjusting the task quantity of each task execution end according to the task execution condition, so that the task quantity of each execution terminal is balanced, and the task completion efficiency is improved.
In this embodiment, the tasks are allocated to the plurality of task execution ends, each task execution end can only check the information of the unmarked data set corresponding to the respective task, and the marked tasks are completed together in a cooperative marking manner, so that not only is the safety of the unmarked data set protected, but also the marking efficiency is improved.
Referring to fig. 4, fig. 4 is a flowchart illustrating a task execution method for defect labeling according to an embodiment of the present invention.
As shown in fig. 4, this includes steps S30 to S31.
Step S30, receiving a task allocation instruction sent by a task sending end, and acquiring unmarked data set information carried by the allocation instruction;
exemplarily, for example, the storage path address of the unlabeled data set 1, the name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, and the name of each picture are obtained, and asd tag information and/or white tag information required for the unlabeled data set 1 are obtained.
And step S31, checking the unmarked data set based on the unmarked data set information, and marking the unmarked data set.
Exemplarily, the unlabeled data set 1 is checked through the storage path address of the unlabeled data set 1 and the name of the unlabeled data set 1, the number of pictures in the unlabeled data set 1, the name of each picture, and asd tag information and white tag information which need to be labeled by the unlabeled data set 1 are checked, and the pictures in the unlabeled data set 1 are labeled through the asd tag information and the white tag information.
In an embodiment, after labeling the unlabeled data set corresponding to the unlabeled data set information based on the labeling information, the method further includes: and feeding back the off-line information so that the task management end can recycle the data set information which is not marked in the task.
Exemplarily, when the task execution end is offline, the task execution end feeds back offline information to notify the task management end. When the task execution end is determined to be offline, the task management end checks the information of the unmarked data set in the task execution end, such as at least one of the name and the number of the unmarked pictures, and recovers the unmarked data set information.
In the embodiment of the invention, the task execution end receives the task allocation instruction, only the unmarked data set information corresponding to the task and the corresponding unmarked data set to be checked can be checked, the unmarked data set is marked, and the unmarked data set information is recycled by the task sending end when the task execution end is off-line, so that the unmarked data set information and the marked task are prevented from being leaked.
Referring to fig. 5, fig. 5 is a schematic block diagram of a computer device according to an embodiment of the present invention. The computer device may be a terminal.
As shown in fig. 5, the computer device includes a processor, a memory, and a network interface connected by a system bus, wherein the memory may include a nonvolatile storage medium and an internal memory.
The non-volatile storage medium may store an operating system and a computer program. The computer program comprises program instructions which, when executed, cause a processor to perform any of the task execution methods for defect labeling.
The processor is used for providing calculation and control capability and supporting the operation of the whole computer equipment.
The internal memory provides an environment for the execution of a computer program on a non-volatile storage medium, which when executed by a processor causes the processor to perform any of the task execution methods for defect labeling.
The network interface is used for network communication, such as sending assigned tasks and the like. Those skilled in the art will appreciate that the configuration shown in fig. 5 is a block diagram of only a portion of the configuration associated with aspects of the present invention and is not intended to limit the computing devices to which aspects of the present invention may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
It should be understood that the Processor may be a Central Processing Unit (CPU), and the Processor may be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, etc. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Wherein, in one embodiment, the processor is configured to execute a computer program stored in the memory to implement the steps of:
determining annotation information, wherein the annotation information at least comprises information of an unmarked data set required by executing a task;
determining authority information so that the task execution end can only view the unmarked data set information corresponding to the distributed tasks;
and sending a task allocation instruction, so that the task execution end checks the unmarked data set according to the unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
In one embodiment, when the processor implements that the task execution end comprises a plurality of tasks, the processor is configured to implement:
sending corresponding tasks to a plurality of task execution ends, and distributing unmarked data set information required by the task execution to the plurality of task execution ends;
and enabling a plurality of task execution ends to cooperatively complete the tasks according to the distributed information of the unmarked data sets.
In one embodiment, when the processor implements that the task execution end comprises a plurality of tasks, the processor is configured to implement: and monitoring the task execution condition of each task execution end in real time, and adjusting the task quantity of each task execution end according to the task execution condition.
In one embodiment, the processor implements that the task execution end comprises a first task execution end; when the task quantity of each task execution end is adjusted according to the task execution condition, the method is used for realizing that:
adjusting the task quantity of each task execution end according to the task execution condition, comprising: and if the fact that the first task execution end does not finish the labeling task of the unmarked data set information distributed when the first task execution end is offline is determined, task recovery is carried out on the labeling task of the unmarked data set information, and the unmarked data set information is distributed to other online task execution ends as a new task, wherein the unmarked data set information comprises the recovered unmarked data set information.
In one embodiment, the processor implements that the task execution end comprises a first task execution end; when the task quantity of each task execution end is adjusted according to the task execution condition, the method is used for realizing that:
if it is determined that the first task execution end completes the labeling task of the distributed unlabeled data set information, recovering the unlabeled data set information which is not labeled by the other task execution ends, and distributing the unlabeled data set information to the first task execution end as a new task, wherein the unlabeled data set information includes the recovered unlabeled data set information.
And displaying at least one of undistributed data set information, data set information of which the task execution end finishes labeling and data set information of which the task execution end does not finish labeling.
In one embodiment, the processor implements a storage path for the unlabeled data set information including at least the unlabeled data set.
Wherein, in one embodiment, the processor is configured to execute a computer program stored in the memory to implement the steps of:
receiving a task allocation instruction sent by a task sending end, and acquiring unmarked data set information carried by the allocation instruction;
and checking the unmarked data set based on the unmarked data set information, and marking the unmarked data set.
In one embodiment, the processor, when implementing after annotating the unlabeled data set, is configured to implement:
and feeding back the off-line information so that the task management end can recycle the data set information which is not marked in the task.
Embodiments of the present invention further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, where the computer program includes program instructions, and when the program instructions are executed, a method implemented by the computer program instructions may refer to various embodiments of the task execution method for defect labeling according to the present invention.
The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the computer device.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments. While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A task execution system for defect labeling, the system comprising:
the task sending end and the task executing end;
the task sending end determines marking information, wherein the marking information at least comprises unmarked data set information required by executing a task and is also used for determining authority information, so that the task executing end can only check the unmarked data set information corresponding to the distributed task and then send a task distribution instruction;
the task execution end receives a task allocation instruction, checks an unmarked data set according to unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
2. A method for task execution for defect labeling, the method comprising:
determining annotation information, wherein the annotation information at least comprises information of an unmarked data set required by executing a task;
determining authority information so that the task execution end can only view the unmarked data set information corresponding to the distributed tasks;
and sending a task allocation instruction, so that the task execution end checks the unmarked data set according to the unmarked data set information carried by the task allocation instruction, and marks the unmarked data set.
3. The task execution method for defect labeling according to claim 2, wherein the task execution end comprises a plurality of task execution ends, including:
sending corresponding tasks to a plurality of task execution ends, and distributing unmarked data set information required by the task execution to the plurality of task execution ends;
and enabling a plurality of task execution ends to cooperatively complete the tasks according to the distributed information of the unmarked data sets.
4. The method for task execution for defect labeling according to claim 2, wherein the task execution ends comprise a plurality of task execution ends, the task execution condition of each task execution end is monitored in real time, and the task amount of each task execution end is adjusted according to the task execution condition.
5. The task execution method for defect labeling of claim 4, wherein the task execution end comprises a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: and if the fact that the first task execution end does not finish the labeling task of the unmarked data set information distributed when the first task execution end is offline is determined, task recovery is carried out on the labeling task of the unmarked data set information, and the unmarked data set information is distributed to other online task execution ends as a new task, wherein the unmarked data set information comprises the recovered unmarked data set information.
6. The task execution method for defect labeling of claim 4, wherein the task execution end comprises a first task execution end; adjusting the task quantity of each task execution end according to the task execution condition, comprising: if the first task execution end is determined to finish the labeling task of the distributed unlabeled data set information, the unlabeled data set information which is not labeled by the other task execution ends is recycled, and the unlabeled data set information is distributed to the first task execution end as a new task, wherein the unlabeled data set information comprises the recycled unlabeled data set information.
7. The method of claim 2, wherein the sending the task allocation instruction is followed by further comprising:
and displaying at least one of undistributed data set information, data set information of which the task execution end finishes labeling and data set information of which the task execution end does not finish labeling.
8. The task execution method for defect labeling of claim 2, wherein the unlabeled dataset information includes at least a storage path of the unlabeled dataset.
9. A method for task execution for defect labeling, the method comprising:
receiving a task allocation instruction sent by a task sending end, and acquiring unmarked data set information carried by the allocation instruction;
and checking the unmarked data set based on the unmarked data set information, and marking the unmarked data set.
10. The method of claim 9, wherein the labeling the unlabeled data set further comprises:
and feeding back the off-line information so that the task management end can recycle the data set information which is not marked in the task.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110778856.1A CN113254221A (en) | 2021-07-09 | 2021-07-09 | Task execution system and method for defect labeling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110778856.1A CN113254221A (en) | 2021-07-09 | 2021-07-09 | Task execution system and method for defect labeling |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113254221A true CN113254221A (en) | 2021-08-13 |
Family
ID=77191085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110778856.1A Pending CN113254221A (en) | 2021-07-09 | 2021-07-09 | Task execution system and method for defect labeling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113254221A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120150859A1 (en) * | 2010-12-10 | 2012-06-14 | Sap Ag | Task-Based Tagging and Classification of Enterprise Resources |
| US20160342624A1 (en) * | 2013-05-01 | 2016-11-24 | Image Searcher, Inc. | Image Tagging System |
| CN110888582A (en) * | 2019-11-25 | 2020-03-17 | Oppo（重庆）智能科技有限公司 | Tag information processing method, device, storage medium and terminal |
| CN111310866A (en) * | 2020-05-09 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Data labeling method, device, system and terminal equipment |
| CN111695613A (en) * | 2020-05-28 | 2020-09-22 | 平安科技（深圳）有限公司 | Data annotation system, computer-readable storage medium, and electronic device |
| CN111859855A (en) * | 2020-06-11 | 2020-10-30 | 第四范式（北京）技术有限公司 | Method, device and equipment for processing labeling task and storage medium |
-
2021
- 2021-07-09 CN CN202110778856.1A patent/CN113254221A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120150859A1 (en) * | 2010-12-10 | 2012-06-14 | Sap Ag | Task-Based Tagging and Classification of Enterprise Resources |
| US20160342624A1 (en) * | 2013-05-01 | 2016-11-24 | Image Searcher, Inc. | Image Tagging System |
| CN110888582A (en) * | 2019-11-25 | 2020-03-17 | Oppo（重庆）智能科技有限公司 | Tag information processing method, device, storage medium and terminal |
| CN111310866A (en) * | 2020-05-09 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Data labeling method, device, system and terminal equipment |
| CN111695613A (en) * | 2020-05-28 | 2020-09-22 | 平安科技（深圳）有限公司 | Data annotation system, computer-readable storage medium, and electronic device |
| CN111859855A (en) * | 2020-06-11 | 2020-10-30 | 第四范式（北京）技术有限公司 | Method, device and equipment for processing labeling task and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022078345A1 (en) | Method for data interaction between multiple devices, and related device | |
| CN107431651B (en) | Life cycle management method and equipment for network service | |
| DE112011105767B4 (en) | network power consumer device | |
| CN114072851A (en) | Image defect identification | |
| CN111130925B (en) | Resource monitoring method and device, computer equipment and storage medium | |
| CN111698283B (en) | Management and control method, device, equipment and storage medium of distributed cluster host | |
| CN110990053A (en) | Method for creating and using machine learning scheme template and device | |
| CN113791765B (en) | Resource arrangement method, device and equipment of cloud service and storage medium | |
| CN106131055A (en) | Method and apparatus for batch operation authority | |
| CN108228618A (en) | The method and apparatus of document verification of data | |
| CN110738038A (en) | Contract text generation method, device, equipment and computer readable storage medium | |
| CN112085019A (en) | Character recognition model generation system, method and device and computer equipment | |
| CN112787853B (en) | Automatic generation method and device of network change scheme and related equipment | |
| CN108427560A (en) | API file management methods, server and storage medium | |
| CN106982131A (en) | Initiate method, the apparatus and system of VNF instantiations | |
| CN108319869A (en) | A kind of method, apparatus opened bid automatically, server and storage medium | |
| CN113254221A (en) | Task execution system and method for defect labeling | |
| CN115002062A (en) | Message processing method, device, equipment and readable storage medium | |
| CN108427557A (en) | A kind of control layout display control method, device and computer readable storage medium | |
| CN113705182A (en) | Contract online generation method and device and readable storage medium | |
| CN114357289A (en) | Information pushing method, information pushing system, electronic device and storage medium | |
| CN114860390B (en) | Container data management method, device, program product, medium and electronic equipment | |
| CN112686391A (en) | Modeling method and device based on federal learning, equipment and storage medium | |
| CN110457273A (en) | A kind of nuclear power station document management method, system and terminal device | |
| CN111125449B (en) | Object information storage method, device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210813 |
|
| RJ01 | Rejection of invention patent application after publication |