Method, system and electronic device for testing effectiveness of data generator

Abstract

The application provides a data generator effectiveness test method, system and electronic equipment. The method comprises: obtaining a plurality of test targets, test backgrounds and a plurality of matching rules for testing; obtaining fourth space-time characteristic data about the matching combination based on physics; forming a first unmatched combination of targets that do not meet the matching rules and a second unmatched combination of backgrounds that do not meet the matching rules; respectively obtaining corresponding first space-time characteristic data and second space-time characteristic data; obtaining third space-time characteristic data through a data generator; comparing the third space-time characteristic data and the fourth space-time characteristic data to obtain a sub-score of the effectiveness of the data generator; obtaining a plurality of sub-scores based on a plurality of matching rules; and forming a total score of the effectiveness of the data generator. The application assists in evaluating whether the data generator can be applied to a required data augmentation scene.

Description

Technical Field

[0001] This invention relates to the field of virtual synthesis validity evaluation technology, and in particular to a method, system, and electronic device for testing the validity of a data generator. Background Technology

[0002] Generating new virtual data based on real data through virtual synthesis has a wide range of applications. For example, a new painting can be virtually synthesized based on a real painting image; another example is to virtually synthesize a new background image based on a real target image.

[0003] The specific objectives of virtual synthesis may differ depending on the specific needs of the application, but they all aim to meet pre-defined requirements. This degree of compliance is referred to as validity in this application. Only when the validity reaches a certain level can the new data from virtual synthesis be applied to a specific task.

[0004] How to evaluate the effectiveness of the generated results of the aforementioned virtual synthesis methods is an issue that urgently needs to be studied.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the general background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a method, system, and electronic device for testing the effectiveness of a data generator.

[0007] This invention provides a method for testing the effectiveness of a data generator. The data generator is capable of synthesizing virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The method includes:

[0008] Obtain multiple test objectives, test backgrounds, and multiple matching rules for testing;

[0009] Based on the matching rules, the test target is placed in the matching test background to form a matching combination, and physical-based fourth spatiotemporal feature data about the matching combination is obtained;

[0010] Based on the matching rules, the test target and the test background are replaced with other test targets and other test backgrounds, respectively, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules;

[0011] Based on the first mismatch combination and the second mismatch combination, the corresponding first spatiotemporal feature data and the second spatiotemporal feature data are obtained respectively;

[0012] Based on the first spatiotemporal feature data and the second spatiotemporal feature data, the third spatiotemporal feature data is obtained through the data generator;

[0013] Compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score for the effectiveness of the data generator;

[0014] Based on the multiple matching rules, multiple sub-scores are obtained;

[0015] Based on the multiple sub-scorings, a total score is formed to assess the effectiveness of the data generator.

[0016] According to the present invention, a method for testing the validity of a data generator is provided, wherein the target and / or the background is composed of multiple objects.

[0017] According to the present invention, a method for testing the validity of a data generator is provided, wherein the test background in the matching combination, the first non-matching combination, and the second non-matching combination includes the test target in the combination in spatiotemporal coordinates.

[0018] According to the present invention, a method for testing the validity of a data generator obtains multiple sub-scores based on multiple matching rules, including:

[0019] For different matching rules, the sub-scores obtained when the corresponding third spatiotemporal feature data and the fourth spatiotemporal feature data have the same degree of similarity may be different.

[0020] According to the present invention, a method for testing the validity of a data generator further includes:

[0021] The test objectives and test backgrounds are classified, and different scores are assigned to different categories.

[0022] Based on the scores assigned to the test background and test target in the matching rules, and the results of comparing the third spatiotemporal feature data and the fourth spatiotemporal feature data, a sub-score of the effectiveness of the data generator is obtained.

[0023] According to the validity testing method of a data generator provided by the present invention, for the matching rule, the more features contained in the test background are the same as the features contained in the test target, the smaller the score corresponding to the matching rule; conversely, the fewer features contained in the test background are the same as the features contained in the test target, the larger the score corresponding to the matching rule.

[0024] According to a method for testing the effectiveness of a data generator provided by the present invention, a total score for the effectiveness of the data generator is formed based on the plurality of sub-scores, including:

[0025] For different categories of test objectives or test backgrounds, obtain multiple sub-scores under each category, and based on the multiple sub-scores under each category, form a total score for the effectiveness of the data generator under each category;

[0026] And / or,

[0027] For different categories of test objectives or test backgrounds, obtain multiple sub-scores for each category, and based on the multiple sub-scores for each category, form a total score for the effectiveness of the data generator across all categories.

[0028] This invention also provides a validity testing system for a data generator. The data generator is capable of synthesizing virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The system includes:

[0029] The acquisition module is used to acquire multiple test targets, test backgrounds, and multiple matching rules for testing purposes.

[0030] The real data acquisition module is used to place the test target in the matching test background based on the matching rules to form a matching combination, and to acquire physical-based fourth spatiotemporal feature data about the matching combination;

[0031] The cross-combination module is used to replace the test target and the test background with other test targets and other test backgrounds respectively, based on the matching rules, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules.

[0032] The input data acquisition module is used to acquire the corresponding first spatiotemporal feature data and second spatiotemporal feature data based on the first mismatch combination and the second mismatch combination, respectively.

[0033] The data generation module is used to obtain the third spatiotemporal feature data based on the first spatiotemporal feature data and the second spatiotemporal feature data through the data generator;

[0034] The comparison module is used to compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score of the effectiveness of the data generator;

[0035] The loop module is used to obtain multiple sub-scores based on multiple matching rules;

[0036] The scoring module is used to form an overall score for the effectiveness of the data generator based on the multiple sub-scorings.

[0037] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the validity testing method for the data generator as described in any of the preceding claims.

[0038] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the validity testing method for the data generator as described in any of the preceding claims.

[0039] The present invention provides a method, system, and electronic device for testing the effectiveness of a data generator, and gives a specific method for scoring and evaluating the effectiveness of the data generator, thereby assisting in assessing whether the data generator can be applied to the required data augmentation scenarios. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0041] Figure 1 A flowchart illustrating a method for testing the validity of a data generator provided by the present invention;

[0042] Figure 2 A schematic diagram of the input and output of a data generator provided by the present invention;

[0043] Figure 3 A schematic diagram of the structure of a data generator validity testing system provided by the present invention;

[0044] Figure 4 This is a schematic diagram of the physical structure of an electronic device provided by the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0046] The effectiveness testing method for the data generator provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application background.

[0047] Figure 1 A flowchart illustrating a validity testing method for a data generator provided by this invention is shown below. Figure 1 As shown, the present invention provides a method for testing the effectiveness of a data generator. The data generator can synthesize virtual third spatiotemporal feature data based on physical first spatiotemporal feature data and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background, the second spatiotemporal feature data includes a second target and a second background, and the third spatiotemporal feature data includes a first target and a second background. The method may include the following steps.

[0048] Preferably, the data generator is an image generator used to synthesize virtual images. For the image, the target can be a specific person or object in the image, and the background can be other people or objects outside the target.

[0049] Preferably, the data generator is an audio generator used to synthesize virtual audio, where the target audio can be the main sound of the audio, and the background is the corresponding noise.

[0050] S100: Obtain multiple test targets, test backgrounds, and multiple matching rules for testing.

[0051] Optionally, the target and / or background may be composed of multiple objects.

[0052] It should be noted that the objectives and backgrounds include not only the aforementioned test objectives and test backgrounds, but also the aforementioned first objective, second objective, first background, and second background.

[0053] Optionally, the test background in the matching combination, the first non-matching combination, and the second non-matching combination includes the test target in the combination in the spatiotemporal coordinates.

[0054] Preferably, in the spatiotemporal coordinates (S O T OTargets with certain characteristics in spacetime coordinates (S) B T B A background with certain characteristics, and the possible prerequisites for forming a combination, are: S B Includes S O T B Includes T O For example, on a time coordinate, a winter target cannot be placed against a summer background; on a spatial coordinate, a target in water cannot be placed against a volcano background.

[0055] S200. Based on the matching rules, the test target is placed in a matching test background to form a matching combination, and the fourth spatiotemporal feature data of the matching combination based on physics is obtained.

[0056] Preferably, a pre-set association table containing matching rules is provided, which includes a test background that matches the test target.

[0057] Specifically, acquiring physical-based fourth spatiotemporal feature data about the matching combination refers to directly acquiring spatiotemporal feature data of the matching combination from physical sensors, such as image information, sound information, olfactory information, etc. of the matching combination.

[0058] S300. Based on the matching rules, the test target and test background are replaced with other test targets and other test backgrounds, respectively, forming a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules.

[0059] S400. Based on the first mismatch combination and the second mismatch combination, obtain the corresponding first spatiotemporal feature data and the second spatiotemporal feature data respectively.

[0060] S500: Based on the first and second spatiotemporal feature data, the third spatiotemporal feature data is obtained through a data generator.

[0061] Preferably, the first and second spatiotemporal feature data are input into the data generator, and the data generator outputs the third spatiotemporal feature data.

[0062] Preferably, the data generator has the ability to automatically separate the target from the background. Figure 2 A schematic diagram of the input and output of a data generator provided by the present invention is shown below. Figure 2 As shown, after separating the target and background from the first and second spatiotemporal feature data, the target and background required by the matching rules are obtained respectively, and the generated data that conforms to the matching rules is synthesized.

[0063] Preferably, the data generator includes a neural network.

[0064] S600. Compare the third and fourth spatiotemporal feature data to obtain a sub-score of the effectiveness of the data generator.

[0065] S700: Based on multiple matching rules, obtain multiple sub-scores.

[0066] It should be noted that, in order to test the effectiveness of the data generator, multiple matching rules were prepared in advance. S700 then traversed all the matching rules and obtained the sub-scores corresponding to all the matching rules.

[0067] Optionally, multiple sub-scores can be obtained based on multiple matching rules, including:

[0068] For different matching rules, the sub-scores obtained when the corresponding third and fourth spatiotemporal feature data have the same degree of similarity may be different.

[0069] It should be noted that different matching rules present varying degrees of difficulty in obtaining generated data with the same level of realism. For more difficult virtual generation, when both have the same validity, the more difficult virtual generation should be given a higher sub-score.

[0070] S800: Based on multiple sub-scores, a total score is formed to assess the effectiveness of the data generator.

[0071] This embodiment provides a specific scoring and evaluation method for the effectiveness of data generators, thereby assisting in assessing whether data generators can be applied to the required data augmentation scenarios.

[0072] Optionally, the method further includes:

[0073] The test objectives and test backgrounds are categorized, and different scores are assigned to each category.

[0074] Based on the scores assigned to the test background and test target in the matching rules, and the results of comparing the third and fourth spatiotemporal feature data, a sub-score of the effectiveness of the data generator is obtained.

[0075] Preferably, for a matching rule, the score of the matching rule can be formed by combining the scores assigned to the test background and the test target; alternatively, the score of the matching rule can be formed by simply considering either the test background or the test target.

[0076] Optionally, for a matching rule, the more features in the test background that are identical to those in the test target, the lower the score of the matching rule; conversely, the fewer features in the test background that are identical to those in the test target, the higher the score of the matching rule.

[0077] It should be noted that the target and the background usually need to have the same temporal or spatial characteristics to be combined, and the higher the proportion of the same characteristics, the lower the difficulty level.

[0078] Optionally, a total score for the effectiveness of the data generator is formed based on multiple sub-scores, including:

[0079] For different categories of test objectives or test backgrounds, obtain multiple sub-scores for each category, and based on the multiple sub-scores for each category, form a total score for the effectiveness of the data generator in each category;

[0080] And / or,

[0081] For different categories of test objectives or test contexts, obtain multiple sub-scores for each category. Based on the multiple sub-scores for each category, form a total score for the effectiveness of the data generator across all categories.

[0082] Preferably, firstly, based on matching rules, the similarity between generated data and real data is determined. Then, based on the similarity and the difficulty of the preset matching rules, sub-scores under those matching rules are obtained. For the same category, multiple preset sub-scores are obtained according to a preset number of times, and finally, the average is taken as the total score for that category.

[0083] Preferably, different numbers of samples are selected for different categories to obtain sub-scores, and all sub-scores are added together to obtain the total score of the test.

[0084] Preferably, the same number of samples are selected from different categories to obtain sub-scores, each category is assigned a weight, and all sub-scores are weighted and summed to obtain the total score of the test.

[0085] Preferably, after the data generator determines that the validity meets the standard, it can use the data generator to arrange and combine a certain number of targets and backgrounds to obtain a large amount of generated data, thereby achieving data augmentation and serving scenarios that require sample expansion, such as scenarios that require a large amount of training data for neural network training, or scenarios that require a large number of samples to be identified in order to test the recognition ability of the intelligent recognition system.

[0086] The validity testing system for the data generator provided by this invention is described below. The validity testing system for the data generator described below can be referred to in correspondence with the validity testing method for the data generator described above.

[0087] Figure 3 This is a schematic diagram of the structure of a data generator validity testing system provided by the present invention, as shown below. Figure 3As shown, this invention also provides a validity testing system for a data generator. The data generator can synthesize virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The system includes:

[0088] The acquisition module is used to acquire multiple test targets, test backgrounds, and multiple matching rules for testing purposes.

[0089] The real data acquisition module is used to place the test target in a matching test background based on the matching rules to form a matching combination and obtain physical-based fourth spatiotemporal feature data about the matching combination.

[0090] The cross-combination module is used to replace the test target and test background with other test targets and test backgrounds respectively based on the matching rules, forming a first mismatch combination where the target does not meet the matching rules and a second mismatch combination where the background does not meet the matching rules.

[0091] The input data acquisition module is used to acquire the corresponding first spatiotemporal feature data and second spatiotemporal feature data based on the first mismatch combination and the second mismatch combination, respectively.

[0092] The data generation module is used to obtain third spatiotemporal feature data based on the first and second spatiotemporal feature data through the data generator;

[0093] The comparison module is used to compare the third and fourth spatiotemporal feature data to obtain a sub-score of the effectiveness of the data generator;

[0094] The loop module is used to obtain multiple sub-scores based on multiple matching rules;

[0095] The scoring module is used to generate an overall score for the effectiveness of the data generator based on multiple sub-scorings.

[0096] This embodiment provides a specific scoring and evaluation method for the effectiveness of data generators, thereby assisting in assessing whether data generators can be applied to the required data augmentation scenarios.

[0097] Figure 4 A schematic diagram of the physical structure of an electronic device provided by the present invention, such as... Figure 4As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a validity testing method of a data generator. The data generator can synthesize virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The method includes:

[0098] Obtain multiple test objectives, test backgrounds, and multiple matching rules for testing;

[0099] Based on the matching rules, the test target is placed in the matching test background to form a matching combination, and physical-based fourth spatiotemporal feature data about the matching combination is obtained;

[0100] Based on the matching rules, the test target and the test background are replaced with other test targets and other test backgrounds, respectively, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules;

[0101] Based on the first mismatch combination and the second mismatch combination, the corresponding first spatiotemporal feature data and the second spatiotemporal feature data are obtained respectively;

[0102] Based on the first spatiotemporal feature data and the second spatiotemporal feature data, the third spatiotemporal feature data is obtained through the data generator;

[0103] Compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score for the effectiveness of the data generator;

[0104] Based on the multiple matching rules, multiple sub-scores are obtained;

[0105] Based on the multiple sub-scorings, a total score is formed to assess the effectiveness of the data generator.

[0106] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer is capable of executing the validity testing method of the data generator provided by the above methods, the data generator being capable of synthesizing virtual third spatiotemporal feature data based on physical first spatiotemporal feature data and second spatiotemporal feature data, wherein the first spatiotemporal feature data includes a first target and a first background, the second spatiotemporal feature data includes a second target and a second background, and the third spatiotemporal feature data includes the first target and the second background, the method comprising:

[0108] Obtain multiple test objectives, test backgrounds, and multiple matching rules for testing;

[0109] Based on the matching rules, the test target is placed in the matching test background to form a matching combination, and physical-based fourth spatiotemporal feature data about the matching combination is obtained;

[0110] Based on the matching rules, the test target and the test background are replaced with other test targets and other test backgrounds, respectively, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules;

[0111] Based on the first mismatch combination and the second mismatch combination, the corresponding first spatiotemporal feature data and the second spatiotemporal feature data are obtained respectively;

[0112] Based on the first spatiotemporal feature data and the second spatiotemporal feature data, the third spatiotemporal feature data is obtained through the data generator;

[0113] Compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score for the effectiveness of the data generator;

[0114] Based on the multiple matching rules, multiple sub-scores are obtained;

[0115] Based on the multiple sub-scorings, a total score is formed to assess the effectiveness of the data generator.

[0116] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a validity testing method for the data generators provided above. The data generator is capable of synthesizing virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data, wherein the first spatiotemporal feature data includes a first target and a first background, the second spatiotemporal feature data includes a second target and a second background, and the third spatiotemporal feature data includes the first target and the second background. The method includes:

[0117] Obtain multiple test objectives, test backgrounds, and multiple matching rules for testing;

[0118] Based on the matching rules, the test target is placed in the matching test background to form a matching combination, and physical-based fourth spatiotemporal feature data about the matching combination is obtained;

[0119] Based on the matching rules, the test target and the test background are replaced with other test targets and other test backgrounds, respectively, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules;

[0120] Based on the first mismatch combination and the second mismatch combination, the corresponding first spatiotemporal feature data and the second spatiotemporal feature data are obtained respectively;

[0121] Based on the first spatiotemporal feature data and the second spatiotemporal feature data, the third spatiotemporal feature data is obtained through the data generator;

[0122] Compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score for the effectiveness of the data generator;

[0123] Based on the multiple matching rules, multiple sub-scores are obtained;

[0124] Based on the multiple sub-scorings, a total score is formed to assess the effectiveness of the data generator.

[0125] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0126] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for testing the validity of a data generator, characterized in that, The data generator can synthesize virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The method includes: Obtain multiple test objectives, test backgrounds, and multiple matching rules for testing; Based on the matching rules, the test target is placed in the matching test background to form a matching combination, and physical-based fourth spatiotemporal feature data about the matching combination is obtained; Based on the matching rules, the test target and the test background are replaced with other test targets and other test backgrounds, respectively, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules; Based on the first mismatch combination and the second mismatch combination, the corresponding first spatiotemporal feature data and the second spatiotemporal feature data are obtained respectively; Based on the first spatiotemporal feature data and the second spatiotemporal feature data, the third spatiotemporal feature data is obtained through the data generator; Compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score for the effectiveness of the data generator; Based on the multiple matching rules, multiple sub-scores are obtained; Based on the multiple sub-scorings, a total score is formed to assess the effectiveness of the data generator.

2. The validity testing method for the data generator according to claim 1, characterized in that, The target and / or the background are composed of multiple objects.

3. The validity testing method for the data generator according to claim 1, characterized in that, The test background in the matching combination, the first non-matching combination, and the second non-matching combination includes the test target in the combination in the spatiotemporal coordinates.

4. The validity testing method for the data generator according to claim 1, characterized in that, Based on the multiple matching rules, multiple sub-scores are obtained, including: For different matching rules, the sub-scores obtained when the corresponding third spatiotemporal feature data and the fourth spatiotemporal feature data have the same degree of similarity may be different.

5. The validity testing method for the data generator according to claim 4, characterized in that, The method further includes: The test objectives and test backgrounds are classified, and different scores are assigned to different categories. Based on the scores assigned to the test background and test target in the matching rules, and the results of comparing the third spatiotemporal feature data and the fourth spatiotemporal feature data, a sub-score of the effectiveness of the data generator is obtained.

6. The method for testing the validity of a data generator according to claim 5, characterized in that, For the matching rule, the more features in the test background that are the same as the features in the test target, the lower the score of the matching rule; conversely, the fewer features in the test background that are the same as the features in the test target, the higher the score of the matching rule.

7. The validity testing method for the data generator according to claim 5, characterized in that, Based on the multiple sub-scorings, a total score for the effectiveness of the data generator is formed, including: For different categories of test objectives or test backgrounds, obtain multiple sub-scores under each category, and based on the multiple sub-scores under each category, form a total score for the effectiveness of the data generator under each category; And / or, For different categories of test objectives or test backgrounds, obtain multiple sub-scores for each category, and based on the multiple sub-scores for each category, form a total score for the effectiveness of the data generator across all categories.

8. A validity testing system for a data generator, characterized in that, The data generator can synthesize virtual third spatiotemporal feature data based on physical first and second spatiotemporal feature data. The first spatiotemporal feature data includes a first target and a first background; the second spatiotemporal feature data includes a second target and a second background; and the third spatiotemporal feature data includes the first target and the second background. The system includes: The acquisition module is used to acquire multiple test targets, test backgrounds, and multiple matching rules for testing purposes. The real data acquisition module is used to place the test target in the matching test background based on the matching rules to form a matching combination, and to acquire physical-based fourth spatiotemporal feature data about the matching combination; The cross-combination module is used to replace the test target and the test background with other test targets and other test backgrounds respectively, based on the matching rules, to form a first mismatch combination where the target does not conform to the matching rules and a second mismatch combination where the background does not conform to the matching rules. The input data acquisition module is used to acquire the corresponding first spatiotemporal feature data and second spatiotemporal feature data based on the first mismatch combination and the second mismatch combination, respectively. The data generation module is used to obtain the third spatiotemporal feature data based on the first spatiotemporal feature data and the second spatiotemporal feature data through the data generator; The comparison module is used to compare the third spatiotemporal feature data and the fourth spatiotemporal feature data to obtain a sub-score of the effectiveness of the data generator; The loop module is used to obtain multiple sub-scores based on multiple matching rules; The scoring module is used to form an overall score for the effectiveness of the data generator based on the multiple sub-scorings.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the validity testing method for the data generator as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the validity testing method for the data generator as described in any one of claims 1-7.

Images