Image annotation method and related equipment

By grouping multiple images to be marked and applying the target image labeling model, the problem of low image labeling efficiency in the prior art is solved, and a more efficient labeling process is achieved.

CN112488139BActive Publication Date: 2024-11-19SHENZHEN INTELLIFUSION TECHNOLOGIES CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201910860363.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-09-11
Publication Date
2024-11-19
Estimated Expiration
2039-09-11

AI Technical Summary

Technical Problem

The existing image annotation method is inefficient and requires multiple images to be marked separately, resulting in waste of resources and low efficiency.

Method used

By grouping multiple images to be marked, determining the target image group, and annotating them using the corresponding target image annotation model, improving the labeling efficiency.

Benefits of technology

Through the application of grouping and model mapping relationships, the efficiency of picture annotation is significantly improved, resource waste is reduced, and labeling quality is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112488139B_ABST
    Figure CN112488139B_ABST
Patent Text Reader

Abstract

The present application discloses a method for image annotation and related equipment, the method comprising: performing an image grouping operation on N to-be-grouped images to obtain M target image groups corresponding to the N to-be-grouped images, where N and M are both integers greater than 1, and M is less than N; determining P target image annotation models corresponding to the M target image groups according to a mapping relationship between the image groups and the image annotation models, where P is an integer greater than 1 and less than or equal to M; performing an image annotation operation on the M target image groups using the P target image annotation models to obtain M target image groups after image annotation. The use of the embodiments of the present application is helpful to improve the image annotation efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of picture annotation, and in particular to a picture annotation method and related equipment. Background Art

[0002] At present, the process of image annotation is as follows: when annotating a picture A among multiple pictures to be annotated, determine the picture annotation model corresponding to the picture A, where the picture A is any one of the multiple pictures to be annotated; use the picture annotation model corresponding to the picture A to annotate to annotate the picture A, and obtain the annotated picture A. This method results in low image annotation efficiency. Summary of the invention

[0003] The embodiments of the present application provide a picture annotation method and related devices for improving the efficiency of picture annotation.

[0004] In a first aspect, an embodiment of the present application provides a method for labeling an image, comprising:

[0005] Performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N;

[0006] Determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between picture groups and picture annotation models, where P is an integer greater than 1 and less than or equal to M;

[0007] The P target picture labeling models are used to perform a picture labeling operation on the M target picture groups to obtain the M target picture groups after picture labeling.

[0008] In a possible example, determining the P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models includes:

[0009] Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups;

[0010] After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model;

[0011] Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1.

[0012] P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

[0013] In a possible example, the using the P target picture labeling models to perform a picture labeling operation on the M target picture groups to obtain the M target picture groups after picture labeling includes:

[0014] Using the j-th target picture annotation model set corresponding to the j-th target picture group, performing a picture annotation operation on the j-th target picture group to obtain the j-th target picture group after picture annotation, wherein the j-th target picture group is any one of the M target picture groups;

[0015] After obtaining the j-th target picture group after the picture annotation, use the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation;

[0016] Until j=M, the M target picture groups after picture annotation are obtained, wherein j is an integer that is incremented with an initial value of 1 and an interval of 1.

[0017] In a second aspect, an embodiment of the present application provides a picture annotation device, including:

[0018] A grouping unit, configured to perform a picture grouping operation on N pictures to be grouped, to obtain M target picture groups corresponding to the N pictures to be grouped, wherein both N and M are integers greater than 1, and M is less than N;

[0019] a determining unit, configured to determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between the picture groups and the picture annotation models, where P is an integer greater than 1 and less than or equal to M;

[0020] The labeling unit is used to perform a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling.

[0021] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing some or all of the steps in the method described in the first aspect of the embodiment of the present application.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which is used to store a computer program. The computer program is executed by a processor to implement some or all of the steps described in the method described in the first aspect of the embodiment of the present application.

[0023] It can be seen that compared to annotating any one of the multiple pictures to be annotated A, using the picture annotation model corresponding to the picture to be annotated A to annotate the picture A to obtain the annotated picture A, in an embodiment of the present application, the picture annotation device performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups, and uses P target picture annotation models corresponding to the M target picture groups to perform a picture annotation operation on the M target picture groups to obtain M target picture groups after picture annotation, which helps to improve the efficiency of picture annotation.

[0024] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background technology, the drawings required for use in the embodiments of the present application or the background technology will be described below.

[0026] Figure 1 It is a schematic diagram of the architecture of a picture annotation system provided in an embodiment of the present application;

[0027] Figure 2A It is a flowchart of a method for marking pictures provided in an embodiment of the present application;

[0028] Figure 2B is a schematic diagram of performing an image annotation operation on M target image groups using P target image annotation models provided by an embodiment of the present application;

[0029] Figure 2C is another schematic diagram of using P target picture annotation models to perform picture annotation operations on M target picture groups provided by an embodiment of the present application;

[0030] Figure 3 It is a flowchart of another image annotation method provided in an embodiment of the present application;

[0031] Figure 4 This is a block diagram of the functional units of a picture annotation device provided in an embodiment of the present application;

[0032] Figure 5 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0033] Specific implementation method

[0034] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0035] Electronic devices may include various handheld devices, vehicle-mounted devices, wearable devices, computing devices or other processing devices connected to a wireless modem with wireless communication functions, as well as various forms of user equipment (UE), mobile stations (MS), terminal devices, etc.

[0036] The embodiments of the present application are described in detail below.

[0037] See also Figure 1 , Figure 1 : is a schematic diagram of the architecture of an image annotation system provided in an embodiment of the present application, the image annotation system includes a processor, wherein:

[0038] A processor is used to perform a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where N and M are both integers greater than 1, and M is less than N;

[0039] The processor is further used to determine P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models, where P is an integer greater than 1 and less than or equal to M;

[0040] The processor is further configured to use the P target image annotation models to perform image annotation operations on the M target image groups to obtain M target image groups after image annotation.

[0041] In a possible example, in performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, the processor is specifically configured to:

[0042] Performing a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types, wherein the N first picture types correspond one-to-one to the N pictures to be grouped; and grouping multiple pictures to be grouped having the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups;

[0043] Alternatively, the second picture type corresponding to each picture to be grouped in N pictures to be grouped is obtained; and multiple pictures to be grouped with the same second picture type in the N pictures to be grouped are divided into the same target picture group to obtain M target picture groups.

[0044] In a possible example, in determining the P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models, the processor is specifically configured to:

[0045] Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups;

[0046] After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model;

[0047] Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1.

[0048] P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

[0049] In a possible example, in terms of performing a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling, the processor is specifically configured to:

[0050] Using the j-th target picture annotation model set corresponding to the j-th target picture group, performing a picture annotation operation on the j-th target picture group to obtain the j-th target picture group after picture annotation, wherein the j-th target picture group is any one of the M target picture groups;

[0051] After obtaining the j-th target picture group after the picture annotation, use the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation;

[0052] Until j=M, the M target picture groups after picture annotation are obtained, wherein j is an integer that is incremented with an initial value of 1 and an interval of 1.

[0053] In a possible example, in terms of performing a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling, the processor is specifically configured to:

[0054] Obtain the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group in the M target picture groups;

[0055] Performing a sorting operation on the M target picture groups according to the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group, to obtain the sorted M target picture groups;

[0056] Using the k-th target picture annotation model set corresponding to the k-th target picture group, perform a picture annotation operation on the k-th target picture group to obtain the k-th target picture group after picture annotation, wherein the k-th target picture group is any one of the M target picture groups after the sorting;

[0057] After obtaining the kth target picture group after the picture annotation, use the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation;

[0058] Until k=M, the M target picture groups after picture annotation are obtained, where k is an initial value of 1 and is an increasing integer with an interval of 1.

[0059] See also Figure 2A , Figure 2A 201 is a flow chart of a method for annotating an image provided by an embodiment of the present application. The method for annotating an image includes steps 201-203, wherein:

[0060] 201: The picture labeling apparatus performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N.

[0061] In a possible example, the picture labeling apparatus performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, including:

[0062] The picture labeling device performs a picture category labeling operation on each of the N pictures to be grouped to obtain N first picture categories, wherein the N first picture categories correspond one-to-one to the N pictures to be grouped;

[0063] The picture labeling device divides multiple pictures to be grouped with the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups.

[0064] Among them, each of the N pictures to be grouped is not labeled with the picture type, and the picture types include person pictures, license plate pictures, animal pictures, etc.

[0065] Specifically, the picture labeling device performs a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types in the following implementation manner:

[0066] The picture labeling device performs a picture type labeling operation on the picture A to be grouped to obtain a first picture type corresponding to the picture A to be grouped, where the picture A to be grouped is any one of the N pictures to be grouped;

[0067] The picture labeling device performs the same operation on (N-1) pictures to be grouped except picture A to be grouped among the N pictures to be grouped, and obtains (N-1) first picture categories, and the (N-1) first picture categories correspond one-to-one to the (N-1) pictures to be grouped.

[0068] Specifically, the picture labeling device performs a picture type labeling operation on the picture A to be grouped, and obtains the first picture type corresponding to the picture A to be grouped in the following manner:

[0069] The picture labeling device determines whether the picture A to be grouped includes at least one of the following: a person, a license plate, or an animal;

[0070] If the picture A to be grouped only includes people, the picture labeling device determines that the first picture type corresponding to the picture A to be grouped is a picture of people;

[0071] If the image A to be grouped only includes a license plate, the image labeling device determines that the first image type corresponding to the image A to be grouped is a license plate image;

[0072] If the to-be-grouped picture A only includes animals, the picture labeling device determines that the first picture type corresponding to the to-be-grouped picture A is an animal picture;

[0073] If the picture A to be grouped includes a person and a license plate, the picture labeling device determines that the first picture type corresponding to the picture A to be grouped is a picture of a person+license plate;

[0074] If the to-be-grouped picture A includes people and animals, the picture labeling device determines that the first picture type corresponding to the to-be-grouped picture A is a picture of people + animals;

[0075] If the to-be-grouped picture A includes a license plate and an animal, the picture labeling device determines that the first picture type corresponding to the to-be-grouped picture A is a license plate + animal picture;

[0076] If the picture A to be grouped includes a person, a license plate and an animal, the picture labeling device determines that the first picture type corresponding to the picture A to be grouped is a picture of a person+license plate+animal.

[0077] The picture labeling device may perform a picture category labeling operation on each of the N pictures to be grouped in a parallel manner to obtain N first picture categories, which will not be described here.

[0078] Specifically, the picture tagging apparatus divides multiple pictures to be grouped with the same first picture type among N pictures to be grouped into the same target picture group to obtain M target picture groups in the following implementation manner:

[0079] The picture labeling device determines whether the first picture type corresponding to the p-th picture to be grouped is the same as the first picture type corresponding to the q-th picture to be grouped, and the p-th picture to be grouped and the q-th picture to be grouped are any one of the N pictures to be grouped;

[0080] If the first picture type corresponding to the p-th picture to be grouped is the same as the first picture type corresponding to the q-th picture to be grouped, the picture labeling device divides the p-th picture to be grouped and the q-th picture to be grouped into the target picture group corresponding to the first picture type corresponding to the p-th picture to be grouped;

[0081] If the first picture type corresponding to the p-th picture to be grouped is different from the first picture type corresponding to the q-th picture to be grouped, the p-th picture to be grouped is divided into the target picture group corresponding to the first picture type corresponding to the p-th picture to be grouped, and the q-th picture to be grouped is divided into the target picture group corresponding to the first picture type corresponding to the q-th picture to be grouped.

[0082] In a possible example, the picture labeling apparatus performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, including:

[0083] The picture labeling device obtains a second picture type corresponding to each picture to be grouped in the N pictures to be grouped;

[0084] The picture labeling device divides multiple pictures to be grouped with the same second picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups.

[0085] Each of the N pictures to be grouped is labeled with a picture type.

[0086] In a possible example, the picture labeling apparatus performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, including:

[0087] The picture labeling device performs a picture category labeling operation on each of the N pictures to be grouped to obtain N1 third picture categories, the N1 third picture categories correspond to the N1 pictures to be grouped one by one, and N1 is less than N;

[0088] The picture annotation device sends a picture category annotation request to the manual annotation platform, which carries (N-N1) pictures to be grouped except N1 pictures to be grouped corresponding to N1 third picture categories among the N pictures to be grouped, and the picture category annotation request is used to instruct the manual annotation platform to feedback the picture category corresponding to each picture to be grouped among the (N-N1) pictures to be grouped;

[0089] The picture annotation device receives (N-N1) third picture categories sent by the manual annotation platform in response to the picture category annotation request, and the (N-N1) third picture categories correspond one-to-one to the (N-N1) pictures to be grouped;

[0090] The picture labeling device divides multiple pictures to be grouped with the same third picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups.

[0091] Among them, a part of the N pictures to be grouped are marked with picture categories, and another part of the pictures to be grouped are not marked with picture categories.

[0092] 202: The picture labeling apparatus determines P target picture labeling models corresponding to the M target picture groups according to a mapping relationship between picture groups and picture labeling models, where P is an integer greater than 1 and less than or equal to M.

[0093] The mapping relationship between the picture group and the picture annotation model is pre-selected and stored in the picture annotation device. The mapping relationship between the picture group and the picture annotation model is shown in Table 1 below:

[0094] Table 1

[0095] Image Group Image annotation model People picture group Character Image Annotation Model License Plate Image Set License Plate Image Annotation Model People + License Plate Picture Group Character image annotation model + license plate image annotation model …… ……

[0096] In a possible example, the picture labeling apparatus determines the P target picture labeling models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture labeling models, including:

[0097] The picture annotation device determines an i-th target picture annotation model set corresponding to the i-th target picture group according to a mapping relationship between the picture groups and the picture annotation models, wherein the i-th target picture group is any one of the M target picture groups;

[0098] After obtaining the i-th target picture annotation model set, the picture annotation device determines the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model;

[0099] Until i=M, the image annotation device obtains M target image annotation model sets, where i is an integer whose initial value is 1 and whose interval is 1 and increases incrementally;

[0100] The image labeling device determines P target image labeling models according to the M target image labeling model sets, and the P target image labeling models are different from each other.

[0101] Specifically, the implementation method of the image labeling apparatus determining P target image labeling models according to the M target image labeling model sets may be:

[0102] The image annotation device selects P target image annotation models from a plurality of target image annotation models included in the M target image annotation model set;

[0103] Alternatively, the picture labeling device divides at least one target picture labeling model with the same target picture labeling model among multiple target picture labeling models included in the M target picture labeling model sets into the same target picture labeling model group to obtain P target picture labeling model groups; the picture labeling device determines P target picture labeling models based on the P target picture labeling model groups.

[0104] Among them, the picture labeling device can determine the P target picture labeling models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture labeling models in a parallel manner, which will not be described here.

[0105] 203: The picture labeling apparatus uses the P target picture labeling models to perform picture labeling operations on the M target picture groups to obtain the M target picture groups after picture labeling.

[0106] The image annotation device uses P target image annotation models in a serial manner to perform image annotation operations on M target image groups to obtain M target image groups after image annotation.

[0107] It can be seen that compared to annotating any one of the multiple pictures to be annotated A, using the picture annotation model corresponding to the picture to be annotated A to annotate the picture A to obtain the annotated picture A, in an embodiment of the present application, the picture annotation device performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups, and uses P target picture annotation models corresponding to the M target picture groups to perform a picture annotation operation on the M target picture groups to obtain M target picture groups after picture annotation, which helps to improve the efficiency of picture annotation.

[0108] In a possible example, the image labeling apparatus uses the P target image labeling models to perform an image labeling operation on the M target image groups to obtain the M target image groups after image labeling, including:

[0109] The picture labeling device uses the j-th target picture labeling model set corresponding to the j-th target picture group to perform a picture labeling operation on the j-th target picture group to obtain the j-th target picture group after picture labeling, wherein the j-th target picture group is any one of the M target picture groups;

[0110] After obtaining the j-th target picture group after the picture annotation, the picture annotation device uses the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation;

[0111] Until j=M, the picture labeling device obtains the M target picture groups after picture labeling, where j is an initial value of 1 and an increasing integer with an interval of 1.

[0112] For example, if Figure 2B As shown, Figure 2B It is a schematic diagram of performing a picture annotation operation on M target picture groups using P target picture annotation models provided in an embodiment of the present application. A picture annotation device uses a person picture annotation model corresponding to a person picture group to perform a picture annotation operation on the person picture group to obtain a person picture group after picture annotation; after obtaining the person picture group after picture annotation, the picture annotation device uses a license plate picture annotation model corresponding to a license plate picture group to perform a picture annotation operation on the license plate picture group to obtain a license plate picture group after picture annotation; after obtaining the license plate picture group after picture annotation, the picture annotation device uses a target picture annotation model set (person picture annotation model + license plate picture annotation model) corresponding to the person + license plate picture group to perform a picture annotation operation on the person + license plate picture group to obtain a person + license plate picture group after picture annotation.

[0113] Specifically, the jth target picture group includes Q pictures to be grouped, where Q is an integer greater than 1. The picture labeling device uses the jth target picture labeling model set corresponding to the jth target picture group to perform a picture labeling operation on the jth target picture group, and the implementation method of obtaining the jth target picture group after picture labeling can be:

[0114] The image annotation device uses the j-th target image annotation model set corresponding to the j-th target image group to perform an image annotation operation on the m-th image to be grouped, and obtains the m-th image to be grouped after the image annotation, and the m-th image to be grouped is any one of the j-th target image group;

[0115] After obtaining the mth picture to be grouped after picture annotation, the jth target picture annotation model set of the picture annotation device performs a picture annotation operation on the (m+1)th picture to be grouped to obtain the (m+1)th picture to be grouped after picture annotation;

[0116] Until m=Q, the picture labeling device obtains Q pictures to be grouped after the pictures are labeled, where m is an integer whose initial value is 1 and increases with an interval of 1;

[0117] The picture labeling device determines the Q pictures to be grouped after the picture labeling as the jth target picture group after the picture labeling.

[0118] Among them, the image labeling device uses the jth target image labeling model set corresponding to the jth target image group in a parallel manner to perform image labeling operations on the jth target image group to obtain the jth target image group after image labeling, which will not be described here.

[0119] In a possible example, the image labeling apparatus uses the P target image labeling models to perform an image labeling operation on the M target image groups to obtain the M target image groups after image labeling, including:

[0120] The picture labeling device obtains the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group in the M target picture groups;

[0121] The picture labeling device performs a sorting operation on the M target picture groups according to the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group, to obtain the sorted M target picture groups;

[0122] The picture labeling device uses the k-th target picture labeling model set corresponding to the k-th target picture group to perform a picture labeling operation on the k-th target picture group to obtain the k-th target picture group after picture labeling, wherein the k-th target picture group is any one of the M target picture groups after the sorting;

[0123] After obtaining the kth target picture group after the picture annotation, the picture annotation device uses the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation;

[0124] Until k=M, the picture labeling device obtains the M target picture groups after picture labeling, where k is an integer that is initially 1 and increases with an interval of 1.

[0125] Specifically, the picture labeling device performs a sorting operation on the M target pictures according to the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group, and obtains the sorted M target picture groups. The implementation method may be: the picture labeling device performs a sorting operation on the M target picture groups according to the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group in descending order of the number of target picture labeling models, to obtain the sorted M target picture groups;

[0126] Alternatively, the picture labeling device performs a sorting operation on the M target picture groups in order of the number of target picture labeling models from small to large according to the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group, to obtain sorted M target picture groups.

[0127] For example, if Figure 2C As shown, Figure 2C It is another schematic diagram of performing picture annotation operations on M target picture groups using P target picture annotation models provided by an embodiment of the present application. The picture annotation device performs a sorting operation on three target picture groups (a person picture group, a license plate picture group, and a person + license plate picture group) in the order of the number of target picture annotation models included in the target picture annotation model set from large to small, and obtains three sorted target picture groups (a person + license plate picture group-a license plate picture group-a person picture group). The picture annotation device uses the target picture annotation model set (a person picture annotation model + a license plate picture annotation model) corresponding to the person + license plate picture group to perform a picture annotation operation on the person + license plate picture group to obtain the person + license plate picture group after picture annotation; after obtaining the person + license plate picture group after picture annotation, the picture annotation device uses the license plate picture annotation model corresponding to the license plate picture group to perform a picture annotation operation on the license plate picture group to obtain the license plate picture group after picture annotation; after obtaining the license plate picture group after picture annotation, the picture annotation device uses the person picture annotation model corresponding to the person picture group to perform a picture annotation operation on the person picture group to obtain the person picture group after picture annotation.

[0128] In a possible example, the image labeling apparatus uses P target image labeling models to perform image labeling operations on M target image groups, and after obtaining the M target image groups after image labeling, the method further includes:

[0129] The image annotation device sends an image annotation request carrying the M target image groups after image annotation to the manual annotation platform, and the image annotation request is used to instruct the manual annotation platform to feed back the M target image groups after secondary image annotation;

[0130] The image annotation device receives M target image groups after secondary image annotation sent by the manual annotation platform in response to the image annotation request.

[0131] It can be seen that compared with the picture labeling device performing picture labeling operations on M target picture groups to obtain M target picture groups after picture labeling, in this example, the picture labeling device sends a picture labeling request carrying the M target picture groups after picture labeling to the manual labeling platform, and receives the M target picture groups after secondary picture labeling sent by the manual labeling platform in response to the picture labeling request, which helps to improve the accuracy of picture labeling.

[0132] See also Figure 3 , Figure 3 3 is a flow chart of another image annotation method provided in an embodiment of the present application. The image annotation method includes steps 301-310, which are as follows:

[0133] 301: The picture labeling apparatus performs a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N.

[0134] 302: The picture labeling apparatus determines an i-th target picture labeling model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture labeling models, wherein the i-th target picture group is any one of the M target picture groups.

[0135] 303: After obtaining the i-th target picture annotation model set, the picture annotation device determines the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture groups and the picture annotation models.

[0136] 304: until i=M, the image labeling apparatus obtains M target image labeling model sets, wherein i is an integer that is initially 1 and increases with an interval of 1.

[0137] 305: The image labeling apparatus determines P target image labeling models according to the M target image labeling model sets, wherein the P target image labeling models are different from each other.

[0138] 306: The picture labeling apparatus obtains the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group in the M target picture groups.

[0139] 307: The picture labeling apparatus performs a sorting operation on the M target picture groups according to the number of target picture labeling models included in the target picture labeling model set corresponding to each target picture group, to obtain the sorted M target picture groups.

[0140] 308: The picture labeling device uses the kth target picture labeling model set corresponding to the kth target picture group to perform a picture labeling operation on the kth target picture group to obtain the kth target picture group after picture labeling, and the kth target picture group is any one of the M target picture groups after the sorting.

[0141] 309: After obtaining the kth target picture group after the picture annotation, the picture annotation device uses the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation.

[0142] 310: until k=M, the picture labeling apparatus obtains the M target picture groups after picture labeling, wherein k is an initial value of 1 and is an increasing integer with an interval of 1.

[0143] It should be noted that Figure 3 The specific implementation process of each step of the method shown can refer to the specific implementation process described in the above method, which will not be described here.

[0144] See also Figure 4 , Figure 4 4 is a functional unit block diagram of a picture annotation device provided in an embodiment of the present application. The picture annotation device 400 includes:

[0145] A grouping unit 401 is configured to perform a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N;

[0146] A determining unit 402 is configured to determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between picture groups and picture annotation models, where P is an integer greater than 1 and less than or equal to M;

[0147] The labeling unit 403 is configured to perform a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling.

[0148] In a possible example, in performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, the grouping unit 401 is specifically configured to:

[0149] Performing a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types, wherein the N first picture types correspond one-to-one to the N pictures to be grouped; and grouping multiple pictures to be grouped having the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups;

[0150] Alternatively, the second picture type corresponding to each picture to be grouped in N pictures to be grouped is obtained; and multiple pictures to be grouped with the same second picture type in the N pictures to be grouped are divided into the same target picture group to obtain M target picture groups.

[0151] In a possible example, in determining the P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models, the determining unit 402 is specifically configured to:

[0152] Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups;

[0153] After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model;

[0154] Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1.

[0155] P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

[0156] In a possible example, in performing a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling, the labeling unit 403 is specifically used to:

[0157] Using the j-th target picture annotation model set corresponding to the j-th target picture group, performing a picture annotation operation on the j-th target picture group to obtain the j-th target picture group after picture annotation, wherein the j-th target picture group is any one of the M target picture groups;

[0158] After obtaining the j-th target picture group after the picture annotation, use the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation;

[0159] Until j=M, the M target picture groups after picture annotation are obtained, wherein j is an integer that is incremented with an initial value of 1 and an interval of 1.

[0160] In a possible example, in performing a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling, the labeling unit 403 is specifically used to:

[0161] Obtain the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group in the M target picture groups;

[0162] Performing a sorting operation on the M target picture groups according to the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group, to obtain the sorted M target picture groups;

[0163] Using the k-th target picture annotation model set corresponding to the k-th target picture group, perform a picture annotation operation on the k-th target picture group to obtain the k-th target picture group after picture annotation, wherein the k-th target picture group is any one of the M target picture groups after the sorting;

[0164] After obtaining the kth target picture group after the picture annotation, use the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation;

[0165] Until k=M, the M target picture groups after picture annotation are obtained, where k is an initial value of 1 and is an increasing integer with an interval of 1.

[0166] See also Figure 5 , Figure 5 1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 500 includes a processor, a memory, a communication interface, and one or more programs. The one or more programs are stored in the memory and configured to be executed by the processor. The program includes instructions for executing the following steps:

[0167] Performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N;

[0168] Determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between picture groups and picture annotation models, where P is an integer greater than 1 and less than or equal to M;

[0169] The P target picture labeling models are used to perform a picture labeling operation on the M target picture groups to obtain the M target picture groups after picture labeling.

[0170] In a possible example, in terms of performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, the above program includes instructions specifically for performing the following steps:

[0171] Performing a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types, wherein the N first picture types correspond one-to-one to the N pictures to be grouped; and grouping multiple pictures to be grouped having the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups;

[0172] Alternatively, the second picture type corresponding to each picture to be grouped in N pictures to be grouped is obtained; and multiple pictures to be grouped with the same second picture type in the N pictures to be grouped are divided into the same target picture group to obtain M target picture groups.

[0173] In a possible example, in terms of determining the P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models, the above program includes instructions specifically for executing the following steps:

[0174] Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups;

[0175] After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model;

[0176] Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1.

[0177] P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

[0178] In a possible example, in terms of using the P target picture annotation models to perform a picture annotation operation on the M target picture groups to obtain the M target picture groups after picture annotation, the program includes instructions specifically for performing the following steps:

[0179] Using the j-th target picture annotation model set corresponding to the j-th target picture group, performing a picture annotation operation on the j-th target picture group to obtain the j-th target picture group after picture annotation, wherein the j-th target picture group is any one of the M target picture groups;

[0180] After obtaining the j-th target picture group after the picture annotation, use the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation;

[0181] Until j=M, the M target picture groups after picture annotation are obtained, wherein j is an integer that is incremented with an initial value of 1 and an interval of 1.

[0182] In a possible example, in terms of using the P target picture annotation models to perform a picture annotation operation on the M target picture groups to obtain the M target picture groups after picture annotation, the program includes instructions specifically for performing the following steps:

[0183] Obtain the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group in the M target picture groups;

[0184] Performing a sorting operation on the M target picture groups according to the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group, to obtain the sorted M target picture groups;

[0185] Using the k-th target picture annotation model set corresponding to the k-th target picture group, perform a picture annotation operation on the k-th target picture group to obtain the k-th target picture group after picture annotation, wherein the k-th target picture group is any one of the M target picture groups after the sorting;

[0186] After obtaining the kth target picture group after the picture annotation, use the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation;

[0187] Until k=M, the M target picture groups after picture annotation are obtained, where k is an initial value of 1 and is an increasing integer with an interval of 1.

[0188] An embodiment of the present application also provides a computer-readable storage medium, which is used to store a computer program. The computer program is executed by a processor to implement part or all of the steps of any method recorded in the above method embodiment. The computer includes an electronic device.

[0189] The embodiment of the present application also provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, the computer program is operable to cause a computer to execute some or all of the steps of any method described in the method embodiment. The computer program product may be a software installation package, and the computer includes an electronic device.

[0190] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0191] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0192] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0193] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0194] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0195] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

[0196] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0197] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for image annotation, It is characterized in that include: Performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, where both N and M are integers greater than 1, and M is less than N; Determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between picture groups and picture annotation models, where P is an integer greater than 1 and less than or equal to M; Using the P target picture annotation models to perform a picture annotation operation on the M target picture groups, to obtain the M target picture groups after picture annotation; The determining, according to the mapping relationship between the picture groups and the picture annotation models, the P target picture annotation models corresponding to the M target picture groups includes: Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups; After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model; Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1. P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

2. The method according to claim 1, It is characterized in that The performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped includes: Performing a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types, wherein the N first picture types correspond one-to-one to the N pictures to be grouped; and grouping multiple pictures to be grouped having the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups; Alternatively, the second picture type corresponding to each picture to be grouped in N pictures to be grouped is obtained; and multiple pictures to be grouped with the same second picture type in the N pictures to be grouped are divided into the same target picture group to obtain M target picture groups.

3. The method according to claim 1 or 2, It is characterized in that The using the P target picture labeling models to perform a picture labeling operation on the M target picture groups to obtain the M target picture groups after picture labeling, comprises: Using the j-th target picture annotation model set corresponding to the j-th target picture group, performing a picture annotation operation on the j-th target picture group to obtain the j-th target picture group after picture annotation, wherein the j-th target picture group is any one of the M target picture groups; After obtaining the j-th target picture group after the picture annotation, use the (j+1)-th target picture annotation model set corresponding to the (j+1)-th target picture group to perform a picture annotation operation on the (j+1)-th target picture group to obtain the (j+1)-th target picture group after the picture annotation; Until j=M, the M target picture groups after picture annotation are obtained, wherein j is an integer that is incremented with an initial value of 1 and an interval of 1.

4. The method according to claim 1 or 2, It is characterized in that The using the P target picture labeling models to perform a picture labeling operation on the M target picture groups to obtain the M target picture groups after picture labeling, comprises: Obtain the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group in the M target picture groups; Performing a sorting operation on the M target picture groups according to the number of target picture annotation models included in the target picture annotation model set corresponding to each target picture group, to obtain the sorted M target picture groups; Using the k-th target picture annotation model set corresponding to the k-th target picture group, perform a picture annotation operation on the k-th target picture group to obtain the k-th target picture group after picture annotation, wherein the k-th target picture group is any one of the M target picture groups after the sorting; After obtaining the kth target picture group after the picture annotation, use the (k+1)th target picture annotation model set corresponding to the (k+1)th target picture group to perform a picture annotation operation on the (k+1)th target picture group to obtain the (k+1)th target picture group after the picture annotation; Until k=M, the M target picture groups after picture annotation are obtained, where k is an initial value of 1 and is an increasing integer with an interval of 1.

5. A picture annotation device, It is characterized in that include: A grouping unit, configured to perform a picture grouping operation on N pictures to be grouped, to obtain M target picture groups corresponding to the N pictures to be grouped, wherein both N and M are integers greater than 1, and M is less than N; a determining unit, configured to determine P target picture annotation models corresponding to the M target picture groups according to a mapping relationship between the picture groups and the picture annotation models, where P is an integer greater than 1 and less than or equal to M; A labeling unit, configured to perform a picture labeling operation on the M target picture groups using the P target picture labeling models to obtain the M target picture groups after picture labeling; In terms of determining the P target picture annotation models corresponding to the M target picture groups according to the mapping relationship between the picture groups and the picture annotation models, the determining unit is specifically used to: Determine an i-th target picture annotation model set corresponding to an i-th target picture group according to a mapping relationship between picture groups and picture annotation models, wherein the i-th target picture group is any one of the M target picture groups; After obtaining the i-th target picture annotation model set, determining the (i+1)-th target picture annotation model set corresponding to the (i+1)-th target picture group according to the mapping relationship between the picture group and the picture annotation model; Until i=M, a set of M target image annotation models is obtained, where i is an integer whose initial value is 1 and whose interval is 1. P target image labeling models are determined according to the M target image labeling model sets, and the P target image labeling models are different from each other.

6. The device according to claim 5, It is characterized in that In performing a picture grouping operation on N pictures to be grouped to obtain M target picture groups corresponding to the N pictures to be grouped, the grouping unit is specifically used to: Performing a picture type labeling operation on each of the N pictures to be grouped to obtain N first picture types, wherein the N first picture types correspond one-to-one to the N pictures to be grouped; and grouping multiple pictures to be grouped having the same first picture type among the N pictures to be grouped into the same target picture group to obtain M target picture groups; Alternatively, obtaining a second picture type corresponding to each of the N pictures to be grouped; A plurality of pictures to be grouped having the same second picture type among the N pictures to be grouped are divided into the same target picture group to obtain M target picture groups.

7. An electronic device, It is characterized in that The method comprises a processor, a memory, a communication interface and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for executing part or all of the steps in the method according to any one of claims 1 to 4.

8. A computer-readable storage medium, It is characterized in that The computer-readable storage medium is used to store a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1 to 4.